Queue item six, and the sixth article this session. The gate blocked the obvious framing again: a third-party-money article scored 4 against the trust reconciliation and travel agency pieces already in this silo. Money held on behalf of others is covered. What is not covered is what happens when a regulator removes the mechanism by which the money was held.
Key Takeaway
Before CARM, an importer could clear goods under its customs broker's Release Prior to Payment security. Since 21 October 2024 it cannot, and the transition period ended at 3 a.m. on 20 May 2025. The broker stopped being a guarantor. That removed a credit risk, a revenue justification, and the single strongest reason a client had not to switch brokers, all in the same instant.
The Verdict, Stated First
Five claims, in descending order of confidence.
One. Importers can no longer use a customs broker's financial security to obtain release prior to payment. CARM became the official system of record on 21 October 2024, and since then importers wanting release before paying duties must post their own security. This is the Canada Border Services Agency's own position.
Two. The transition ended at 3 a.m. EDT on 20 May 2025, after a 180-day period extended by thirty days. Importers who had not posted security by then lost release prior to payment.
Three. On our own arithmetic the gross and net presentations of the same brokerage differ by 43.9 times on the revenue line and by nothing at all on profit. Revenue of $184.2 million against revenue of $4.2 million, and $1,050,000 of operating profit either way.
Four. CARM weakened the case for the larger number. A broker fronting its own bond bore a real credit risk, which is one of the classic indicators pointing toward principal treatment. That risk is gone. This is our reading and no source we found frames it this way.
Five. The bond was also the commercial moat. A client whose goods cleared under the broker's security had a switching cost that has now disappeared. Least confident of the five, because it is an inference about competitive dynamics that we could not test against any data on broker churn.
Our Grades For These Claims
We grade our own sourcing before anyone else has to.
Claims one and two are well sourced and come from the CBSA itself. We read Customs Notice 24-27 on the CARM October implementation transition measures, and two CBSA news releases, one of 28 February 2025 reminding importers to post security and one of 26 March 2025 granting the thirty-day extension. The dates and the substance are the agency's own.
Claim three is arithmetic on invented numbers. The $180 million of duty and tax, the $4.2 million of fees and the $3.15 million of operating cost are our assumptions. The 43.9 times ratio follows from them and moves with them. What does not move is that the two presentations produce identical profit.
Claim four is ours and is inference. We are reasoning from a change in who bears credit risk to a change in a revenue recognition analysis. That is a defensible chain and it is a chain, not a citation.
We did not obtain the revenue recognition guidance. Not the principal versus agent indicators under IFRS 15, not the ASPE equivalent. We describe the shape of the test from general knowledge and we do not cite a standard, because citing a standard we have not read would be worse than admitting we have not read it.
The Customs Act section 17 amendment is single-sourced and secondary. We report it because it is material and recent. It is the weakest thing in this article.
A Note On Method
What we obtained: CBSA Customs Notice 24-27 on CARM October implementation transition measures; CBSA news release of 28 February 2025 on the financial security requirement; CBSA news release of 26 March 2025 granting the thirty-day extension to 20 May 2025; and a range of customs broker, carrier and surety commentary describing the practical effect on importers and brokers.
What we did NOT obtain:
- Any accounting standard. We frame the principal versus agent question and describe the kind of indicators that bear on it. We do not cite IFRS 15, ASPE or any specific paragraph, because we did not read them for this article.
- Customs Notice 25-32, which a secondary source cites for amendments to section 17 of the Customs Act effective 1 January 2026. We report that at second hand and flag it every time.
- Memorandum D17-1-8 on applying for a surety bond, which one source references.
- Any broker's financial statements, fee schedule, bond amount or client count.
- Any data on broker switching or client churn before or after CARM. Our moat argument is unsupported by evidence and we say so where we make it.
One boundary. This article is about customs brokerage. Freight forwarding raises a related but distinct principal versus agent question, which we treat briefly and separately near the end rather than blending the two, because the facts that decide them are different.
The Question That Changes Revenue Forty-Fold
A customs broker files declarations and pays duty and tax to the Crown on an importer's behalf. The importer reimburses it. The broker charges a fee for the work.
So there are two flows. A very large one that passes through, and a small one the broker keeps. The accounting question is whether the large one is the broker's revenue.
If the broker is acting as principal, it is buying and reselling a service, and gross revenue includes the duty. If it is acting as agent, it is arranging for something to be provided by another party, and revenue is the fee alone.
This is the cross-cutting question the industry map flags as living in wholesale, advertising, travel, freight forwarding, e-commerce marketplaces, collection agencies and insurance. It is the same question everywhere and it has an unusual property.
It changes reported revenue by multiples and profit by nothing.
That combination is what makes it dangerous. An error in cost allocation changes profit and somebody notices. An error here leaves profit exactly where it was and moves the top line by a factor that would be a scandal anywhere else.
What Gross And Net Look Like
Ours, on an invented brokerage. Assume it clears $180,000,000 of duty and tax a year for its clients, charges $4,200,000 in fees, and has $3,150,000 of operating cost.
Gross presentation. Revenue $184,200,000. Cost of sales $180,000,000. Gross profit $4,200,000. Operating cost $3,150,000. Operating profit $1,050,000. Operating margin 0.570 percent.
Net presentation. Revenue $4,200,000. Operating cost $3,150,000. Operating profit $1,050,000. Operating margin 25.0 percent.
The revenue line differs by 43.9 times. The profit is identical to the dollar. The margin differs by a factor of 44 and describes the same business on the same day.
Neither presentation is a lie. One of them is correct and the other is not, and which is which depends on a judgment about the substance of the arrangement rather than on any number in the ledger.
Why The Same Profit Is Not The Same Business
It would be reasonable to ask why this matters if profit is identical. Four reasons, all practical.
Valuation. Businesses in this sector are frequently discussed in revenue multiples. At a 1.2 times revenue multiple, our invented broker is worth $221,040,000 on the gross presentation and $5,040,000 on the net presentation. The correct answer is somewhere around the net figure, because the multiple was derived from businesses whose revenue means something. Applying a revenue multiple to a gross-presented broker produces a number that is absurd by two orders of magnitude, and the absurdity is not always obvious to a buyer who has not thought about it.
Covenants. A bank covenant expressed against revenue, or a ratio with revenue in the denominator, behaves completely differently under the two presentations. A change in presentation with no change in the business can breach or cure a covenant.
Margin comparisons are meaningless across the boundary. A 0.570 percent operating margin and a 25 percent operating margin are the same business. Any benchmarking exercise that mixes gross-presenting and net-presenting brokers is comparing nothing to nothing.
Growth looks different. Under gross presentation, revenue moves with duty rates and commodity prices. A tariff increase raises the broker's reported revenue without the broker doing anything, and a tariff cut reduces it. Under net presentation, revenue moves with the number and complexity of transactions, which is what the broker actually sells.
Ours. That last point deserves emphasis in the current environment. A gross-presenting broker in a period of tariff volatility will report revenue swings that have no connection to its own performance, and everyone reading those statements will attribute them to the business.
Most Of It Is Not Duty
One composition point that changes how large the pass-through really is, and that we have been glossing over.
A large part of what a customs broker remits on an importer's behalf is not customs duty at all. It is GST on imported goods. On many commodities, particularly those entering from free trade partners at zero duty, the GST is the overwhelming majority of what is paid at the border.
That matters for two reasons.
The GST is generally recoverable by the importer. A registrant importing for use in commercial activity claims the tax paid at the border as an input tax credit. So the economic cost of that component to the importer is not the amount, it is the timing: cash out at the border and cash back on the return.
Which makes the agency analysis on that component even clearer. The broker is remitting, on behalf of a registrant, an amount that the registrant will recover from the same Crown. It is difficult to construct any account of that flow in which the broker controls anything.
Ours, and worth stating as a caution about our own arithmetic. When we said our invented broker clears $180,000,000, we described it as duty and tax and did not split it. If the split is heavily weighted to GST, then a great deal of the gross revenue figure a gross-presenting broker reports is recoverable consumption tax passing through on its way back to the person who paid it.
Presenting that as revenue is a strong claim. We do not think it survives contact with the question of what the broker actually sold.
What Decides It
We are going to describe the shape of the test without citing a standard, because we did not read one for this article and pretending otherwise would be worse than the gap.
The question is always whether the entity controls the specified good or service before it is transferred to the customer, and the analysis is supported by indicators [6]. The ones that recur are: who is primarily responsible for fulfilling the promise, who bears inventory or equivalent risk, and who has discretion in establishing the price.
Run a customs broker through them.
Primary responsibility. The duty is a statutory obligation of the importer. The broker prepares and files. On the face of it the importer owes the Crown, not the broker.
Discretion in pricing. None whatever. The duty is what the tariff says it is. The broker sets its fee and nothing else.
Risk. This is where it used to get interesting, and it is the reason this article exists.
Two of the three indicators have always pointed clearly toward agent. The third, before CARM, did not point so clearly, because of the bond.
The Bond Was The Answer To One Of Them
Under the old arrangement, an importer could obtain release of goods before paying duties and taxes by relying on its customs broker's Release Prior to Payment security. The broker carried a blanket bond and its clients cleared under it [4].
Think about what that bond actually was. It was a guarantee, given by the broker to the Crown, that duty owing on its clients' importations would be paid. If a client failed, the security stood behind the obligation.
So the broker bore a real credit risk on amounts that were not its own debt. It fronted, or stood behind, other people's statutory obligations at a scale enormously larger than its own revenue.
Ours, on the invented broker. At $180,000,000 of annual duty and tax and a peak month at 1.35 times the average, peak monthly payables are of the order of $20,250,000. That is the scale of exposure standing behind a business earning $1,050,000 of operating profit.
Nineteen times operating profit, in contingent exposure, on obligations the broker did not owe.
That is a genuine risk position and it is the one indicator that gave a gross presentation anything to stand on. A party bearing multi-million dollar credit risk on a transaction is not obviously a pure arranger.
We are not saying gross was right. We are saying the argument existed, and that it rested almost entirely on the bond.
What CARM Did
The CBSA Assessment and Revenue Management system replaced the legacy import accounting system, and it changed the financial security model completely.
Since CARM became the official system of record on 21 October 2024, importers have not been able to use a broker's Release Prior to Payment security to obtain release of goods before paying duties [1]. Importers wanting release prior to payment must post their own financial security through the CARM Client Portal.
The CBSA's own language is direct: a new financial security model exists whereby all commercial importers must post their financial security in CARM to benefit from the programme, and they can no longer use the financial security posted by their customs brokers [2].
Security can be posted either as a deposit into the importer's account or through a financial security agreement with a provider, which in practice means a surety bond.
The broker can still act for the importer once authority is delegated in CARM, and brokers routinely arrange bonds on their clients' behalf. But the security is the importer's, the account is the importer's, and the broker cannot create the importer's CARM account for it.
So the broker went from guarantor to arranger. The service did not change. The risk did.
The Dates That Matter
The transition was extended once and the dates are worth having precisely, because they determine which financial year the change lands in.
21 October 2024. CARM becomes the official system of record. From this date importers cannot use a broker's security.
180-day transition period commences, allowing importers time to obtain their own security while still benefiting from release prior to payment.
28 February 2025. CBSA issues a reminder to commercial importers to submit financial security before the transition measure ends in April.
26 March 2025. In response to stakeholder feedback, CBSA grants a thirty-day extension.
3 a.m. EDT, 20 May 2025. The transition period concludes. As of that moment, all importers requiring release prior to payment must have posted their own financial security. Those who had not were no longer eligible [3].
For a broker with a December year end, that places the substantive change in the 2025 financial year with the announcement in the 2024 one. For a broker with a March or June year end it straddles differently.
Ours. The practical point is that any year-over-year comparison spanning that date is comparing two different businesses. If the broker previously recognised anything in respect of the bond facility, whether a fee, a provision, or a contingent liability disclosure, the comparative will not be like for like and the variance analysis will show a movement that no operational explanation fits.
What Release Prior To Payment Is Actually Worth
The programme at the centre of all this is a working capital instrument, and it is worth pricing so that the stakes in the transition are visible.
Release prior to payment lets goods be released before duties and taxes are paid, with accounting and payment deferred to a monthly settlement. That is float, and float has a value.
Ours, on an invented importer paying $2,400,000 a year in duty and tax, which is $6,575 a day. At a nine percent cost of capital:
- 30 days of float: working capital of $197,260, costing $17,753 a year
- 45 days: $295,890, costing $26,630
- 60 days: $394,521, costing $35,507
Now the cost of keeping it. Security is sized against highest monthly payables, which on our importer at 1.35 times the average month is $270,000.
- A surety bond at a 1.5 percent premium costs $4,050 a year
- A cash deposit ties up the full $270,000, costing $24,300 a year in opportunity cost at the same nine percent
Two conclusions follow, both ours.
Release prior to payment is worth roughly six and a half times what a bond costs to keep it. At 45 days of float the benefit is $26,630 against a $4,050 premium. Losing it on 20 May 2025 for want of a bond was, on these numbers, an expensive administrative failure.
Cash security costs about six times what a bond costs. $24,300 against $4,050. An importer that posted cash because it was the quicker route through the portal in May 2025 is paying a substantial annual premium for that convenience and may not have revisited it since.
Both of those numbers scale directly with the importer. They are the sort of thing a broker could compute for every client on its book in an afternoon, and it is a better conversation than most of what fills the relationship.
The Bond Was Also The Moat
Now the commercial consequence, which we think is larger than the accounting one and which we can support least.
Consider what switching brokers used to involve for an importer clearing under its broker's security. The client's ability to get goods across the border without paying duty up front was a function of that specific broker's bond. Moving meant arranging the same accommodation elsewhere, and the incumbent broker knew it.
That is a switching cost, and switching costs are what keep low-margin service relationships in place. On our invented broker, at 25 percent net margin on fees, the business is not held together by being cheap.
Under CARM the importer has its own account, its own security, and a delegation of authority to the broker that it granted and can withdraw. The financial dependency is gone. What remains is the service, on its merits, priced against competitors who can be given the same delegation tomorrow.
We want to be careful here because we have no evidence. We did not find any data on broker switching rates before or after CARM, and we looked. The argument is structural and it is the kind of structural argument that is often right and occasionally completely wrong, because relationships persist for reasons that do not show up in an incentive analysis.
What we can say without the data: if a broker's client retention was partly a function of the bond, that function ended on 20 May 2025, and a retention rate that has not yet moved is not evidence that it will not.
What Left The Balance Sheet
The change had a balance sheet effect in both directions and it is worth separating them, because one is good news and one is not.
What went away. The blanket bond and its associated contingent exposure. Whatever surety premium the broker was paying to maintain a facility sized against its whole client base. Whatever provision or disclosure it carried for the possibility of a client defaulting on duty the broker had guaranteed. And, in most arrangements, a working capital requirement, because a broker fronting duty is financing the gap between paying the Crown and being reimbursed by the client.
That last one is real money. A broker paying duty on Monday and collecting on the following month's terms carries a receivable that is a multiple of its own revenue, funded somehow.
What did not go away. The credit risk on the broker's own fees, which is small. And, in practice, the client relationship management burden of getting several hundred importers registered, secured and delegated inside a deadline, which was a real and largely unbillable cost through late 2024 and the first half of 2025.
Ours, and the point of the section. A broker that came out of this with a smaller balance sheet, a smaller surety line, and less working capital tied up should show a better return on capital and a worse absolute profit if the accommodation was priced into the fee. Those two movements in the same year, in opposite directions, are exactly the configuration in which a management report tells a confusing story. Anyone reading a 2025 brokerage result needs to know which effects are in it.
And What Arrived On 1 January 2026
One further development and we flag it as our weakest fact before stating it.
A secondary source reports that amendments to section 17 of the Customs Act took effect on 1 January 2026, with the effect that the importer of record is jointly liable for duties, taxes and post-accounting adjustments, and that the entity named on the accounting documents owns the debt. It cites Customs Notice 25-32 [5].
We did not obtain Customs Notice 25-32 and we did not read the amendment. This comes to us from a single logistics company's blog post, which is several rungs below the sourcing we have used for everything else in this article. We report it because if it is accurate it matters a great deal, and because a reader who is going to act on this article should know that this particular thread exists and needs pulling.
If accurate, the direction is consistent with everything else CARM did: liability follows the party named on the accounting documents, which is the importer, rather than resting anywhere it can be diffused.
For a broker the question that follows is who exactly is named. A broker that is shown as importer of record on any transaction, which happens in some arrangements, would on that reading own the debt. That is not a theoretical point and it is the single item in this article we would most want a broker to verify against the actual notice rather than against our account of somebody else's account of it.
A correction to our own research, added after publication. We wrote this section as though the importer-of-record point were new ground for this programme. It is not. An article already published here, on classification, valuation and origin under CARM, makes the substantive point directly and better: the importer of record bears legal responsibility for the accuracy of a declaration, the broker acts as agent on the information supplied, and the common failure is that the broker classifies using the importer's own product description while neither party regards itself as having made the determination. That treatment does not rest on the single secondary source we relied on above. A reader who wants this point should go there rather than to us. The link is in the related reading below.
We record the miss rather than quietly adding the link, because the reason for it is worth knowing. Our topic gate reads index cards, not article bodies, deliberately, because bodies carry incidental vocabulary that produces false positives. The cost of that design is exactly this case: the existing article's card describes classification, valuation and origin, and says nothing about security or importer liability, so two substantive sections of it were invisible to the gate. The gate scored this topic clear and was working as designed. The design has a blind spot.
CARM Changed The Revenue Question
Now put the two halves together, and this section is entirely our reasoning.
The principal versus agent analysis for a customs broker always had two indicators pointing firmly at agent. The importer is primarily responsible for a statutory obligation. The broker has no discretion over the amount of duty whatsoever.
The third indicator, risk, was the only one with anything to say for the other side, and what it had to say was the bond. A broker standing behind $20 million of peak monthly exposure on obligations it did not owe was bearing something.
CARM removed it. As of 20 May 2025 the broker guarantees nothing.
So the analysis that may have been arguable before is, on our reading, considerably less arguable now. A broker that presented gross partly on the strength of the credit risk it bore has lost the fact the argument rested on.
We want to be precise about the limits of that. We are not saying every gross presentation was wrong before or is wrong now. Arrangements differ, some brokers do things that go beyond arranging, and the analysis is done on a contract rather than on an industry. We did not read the standard and we have not seen anyone's arrangement.
What we are saying is narrower and, we think, useful: a specific fact that supported one side of a finely balanced judgment ceased to exist on a known date, and any conclusion reached before that date was reached on facts that no longer hold. A brokerage that documented its principal versus agent conclusion in 2019 and has not revisited it has a conclusion resting on a bond it no longer posts.
What Should Have Happened To The Statements
If the analysis changed, something should have happened in the accounts, and the mechanics of it are not trivial.
A change from gross to net presentation is not a small reclassification. On our invented broker it takes revenue from $184.2 million to $4.2 million, a reduction of 97.7 percent, with no change in profit.
Three consequences follow immediately.
Comparatives. A restatement of the comparative period would show the prior year at $4.2 million too, which is coherent but startling. Presenting the current year at $4.2 million against a comparative at $184.2 million would show a 97.7 percent revenue decline in a business that grew.
Covenants. Anything measured against revenue moves by a factor of 44. This needs to be raised with a lender before it appears in a statement, not after.
Explanation. A reader of the statements who does not understand why the top line moved by two orders of magnitude will assume something catastrophic happened.
We are not going to say whether such a change should be treated as a change in accounting policy, a correction, or a change arising from a change in facts. Those are different things with different treatments and the answer depends on why the earlier conclusion was reached, which we cannot know. We did not obtain the guidance and we are not going to reason from a customs notice to a presentation conclusion. That is the move that produces confident wrong answers and we made a version of it earlier this session on the distillery article and published the correction.
The Freight Side Is A Different Question
The queue entry paired customs brokerage with freight forwarding, and many businesses do both. They should not be analysed together, because the facts that decide them differ completely.
In customs brokerage, the passed-through amount is a statutory duty. There is no supplier, no negotiation, no margin available on it, and no discretion in the price. That is about as clean an agency fact pattern as exists.
In freight forwarding, the passed-through amount is the price of transport bought from a carrier. That is an ordinary commercial supply, negotiated, with a price the forwarder may influence and on which a spread may be earned.
Those are not the same. A forwarder that buys capacity at one price and sells it at another is doing something a customs broker filing a declaration is not doing, and the analysis has to reflect that.
The distinction we would draw, ours. The question is whether the forwarder is arranging transport that the carrier provides to the shipper, or is itself undertaking to carry and buying the capacity to do so. The first is arrangement. The second is not, whatever the invoice says.
This matters because a business doing both may reach different answers on its two revenue streams, and having different answers is correct rather than inconsistent. A single blended presentation across both would be the error.
Issuing Your Own Document
There is a factual test in the freight world that we think is more useful than most of the abstract discussion, and it is whose document the shipper holds.
A forwarder that arranges carriage and hands the shipper the carrier's document has arranged something. A forwarder that issues its own transport document, in its own name, has undertaken to carry the goods and has bought carriage from an actual carrier in order to perform that undertaking.
The second is not an arrangement. The forwarder is primarily responsible to the shipper for performance, it has a legal obligation running to the shipper regardless of what the carrier does, and the spread between what it charged and what it paid is its own.
Ours, and offered as a practical first question rather than as a rule. Whose name is on the document the customer holds is a fact anyone in the business can establish in an afternoon, and it aligns with the substance far more often than the invoicing convention does. Businesses that bill gross because that is how the software was configured, or net because that is how the previous controller did it, are answering a question of substance with a matter of habit.
We did not obtain any accounting guidance on this and we have not looked at transport law. The test above is our own framing of a distinction that we believe is widely understood in the industry, and a business relying on it should get a proper analysis rather than a heuristic from an article.
The Business That Does Both
Most Canadian trade services businesses of any size do customs brokerage, freight forwarding, and often warehousing and distribution, and they frequently issue one invoice.
That single invoice may contain duty and GST remitted to the Crown, ocean or air freight bought from a carrier, terminal and port charges, the forwarder's own handling, and the brokerage fee.
On the analysis above those lines do not all behave the same way. The duty is about as clearly an agency flow as exists. The freight depends on whose document the shipper holds. The handling is the business's own service.
So a correct presentation may be net on one component and gross on another, inside a single invoice. That is not inconsistency. It is the analysis being applied to different facts.
The practical difficulty is that most billing systems do not carry the distinction, because the distinction is not needed to raise the invoice or to collect it. It is needed only to report. So the data required to present correctly frequently does not exist in a retrievable form, and reconstructing it after the fact means going line by line through a year of invoices.
Ours. This is the reason we would expect the wrong answer to persist in this sector even where somebody knows it is wrong. The cost of fixing it is a systems project, the benefit is a presentation change that leaves profit untouched, and nobody's bonus depends on it.
If You Run A Brokerage Or Forwarder
Five things, in the order we would look at them.
Find out when your principal versus agent conclusion was last documented. If it predates 21 October 2024, it was reached on facts that included a bond you no longer post for clients.
Work out what the bond was worth to you, in both directions. Surety premium, working capital, contingent exposure, and whatever the accommodation contributed to your fee. Two of those improved and one may have disappeared.
Talk to your lender before your statements do. If presentation changes, anything measured against revenue moves by a factor that will look like a catastrophe on a covenant certificate.
Check whether you are ever named as importer of record. On the section 17 point we report at second hand, the entity on the accounting documents owns the debt. Verify that against the actual notice.
Separate your revenue streams before you analyse them. Duty, purchased freight and your own services are three different questions and one blended answer will be wrong on at least two of them.
If You Advise One
Four checks we would run on any customs brokerage or forwarding engagement.
Whether revenue is presented gross or net, and whether anyone can say why. The second half of that question is the one that finds problems. A presentation that exists because the system was set up that way is not a conclusion.
Whether the 2025 comparatives are like for like. The transition ended on 20 May 2025. Any year spanning that date contains two different risk structures, and if the broker previously earned or provided anything in respect of the bond, the comparative will not tie to an operational explanation.
Whether the billing system distinguishes disbursements from services. If it does not, a correct presentation cannot be produced without a manual exercise, and you should know that before you need the number rather than during the audit.
Whether working capital fell in 2025 and nobody explained why. A broker no longer fronting duty should need less funding. If the facility is unchanged, either the arrangement did not work the way we have described or there is headroom nobody has noticed.
And one thing to resist. Do not conclude that gross presentation was always wrong. The bond was a real risk position, arrangements differ, and the analysis is done on a contract. What changed is a fact, not a rule.
What To Do
If you take one thing from this article, take the date. On 20 May 2025 customs brokers stopped guaranteeing their clients' duty obligations to the Crown. That was a change in risk, in working capital, and in the commercial relationship, and it happened to an entire industry on the same morning.
If you take two, take the forty-fold. The gross and net presentations of the same brokerage differ by 43.9 times on our invented numbers and by nothing at all on profit. That is the largest number in the business and it is decided by a judgment rather than by a measurement.
If you are advising a broker or forwarder this quarter, the highest-value single question is when the principal versus agent conclusion was last written down. If the answer is before October 2024, or if there is no answer, the work has not been done on current facts.
The Limits Of This Analysis
Long and specific, because a limits section that is short is decoration.
We did not read any accounting standard for this article. Not IFRS 15, not the ASPE equivalent. We describe the shape of the principal versus agent analysis from general knowledge and deliberately cite no paragraph. Anything we say about the test should be checked against the standard by someone who has it open.
The section 17 amendment is single-sourced from a logistics company blog. We did not obtain Customs Notice 25-32 or the amendment. It is the weakest fact here and we flagged it in the body as well.
Every number about the brokerage is invented. $180 million of duty, $4.2 million of fees, $3.15 million of operating cost, and the 1.35 times peak month. The 43.9 times ratio and the $20.25 million bond exposure follow from them. What does not depend on them is that the two presentations produce identical profit.
So is the importer. The $2.4 million of annual duty and tax, the nine percent cost of capital, the 30 to 60 day float range and the 1.5 percent bond premium are all our assumptions. The two ratios they produce, release prior to payment being worth about six and a half times a bond premium and cash security costing about six times a bond, are arithmetic on those assumptions. The bond premium in particular is a rate we did not source and which will vary a great deal by importer credit quality.
We did not split duty from GST in the $180 million figure, and the split matters to how strong the agency argument is on the pass-through. We raised the point and did not quantify it, because we have no basis for choosing a split.
The moat argument has no evidence behind it. We found no data on broker switching rates before or after CARM and we looked. It is structural reasoning about incentives and those are often right and sometimes badly wrong.
We do not know how brokers actually priced the bond accommodation. We assume it contributed something to the fee, which is why we treat its loss as having a revenue effect. If it was provided free as a relationship service, the loss is pure risk reduction with no revenue consequence and half of our balance sheet section falls away.
We did not examine any real arrangement. Brokerage contracts differ, some brokers do more than arrange, and the analysis is contract by contract.
The freight forwarding half is thinner than the brokerage half, deliberately, because it is a different question and we treated it as a distinction to be drawn rather than a subject to be covered. The document test we offer is our own framing.
Nothing here addresses the Canadian trade environment more broadly, and tariff volatility bears directly on gross-presented revenue in a way we mention once and do not develop.
Nothing here is advice on a particular business, and the CARM position is current only as at the date in the meta bar.
Frequently Asked Questions
Can an importer still clear goods under its customs broker's bond?
What happens to an importer without its own financial security?
Why does gross versus net presentation matter if profit is the same?
Did CARM change the accounting analysis?
How large was the risk a broker used to carry?
Is freight forwarding the same question?
What is the practical test for a freight forwarder?
References
- Canada Border Services Agency, Customs Notice 24-27, CARM October implementation transition measures. Source of the statements that CARM became the official system of record effective 21 October 2024; that since that date importers have not been able to use a broker's Release Prior to Payment security to obtain release before paying duties; that importers wanting release prior to payment must post their own financial security through the CARM Client Portal; that a 180-day transition period was provided and extended to 20 May 2025 and is now concluded; that as of 20 May 2025 importers who have not posted security are not eligible for release prior to payment; and that use of a broker BN15 was permitted for a 12-month transition under specified scenarios. Note: the agency's own customs notice and the strongest source in this article. CBSA
- Canada Border Services Agency news release, 28 February 2025, reminding importers to submit financial security before the CARM transition measure ends. Source of the statement that a new financial security model exists whereby all commercial importers must post their financial security in CARM and can no longer use the financial security posted by their customs brokers, and of the description of the consequences of not doing so, being loss of electronic release before payment and longer paper-based processing. Note: a CBSA news release, primary as to the agency's position. CBSA
- Canada Border Services Agency news release, 26 March 2025, granting commercial importers a 30-day extension to submit financial security. Source of the extension to 3 a.m. EDT on 20 May 2025, of the confirmation that security may be posted either by deposit into the importer account or by entering a financial security agreement with a provider, and of the stated benefits of the programme including accelerated release and deferred accounting and payment. Note: primary, and the source of the exact deadline. CBSA
- Customs broker, carrier and surety commentary on CARM, including material from Livingston International, King Bros. Limited, FedEx Canada and surety bond providers. Used for the practical description of the pre-CARM arrangement in which a broker carried a blanket bond covering all of its clients; for the statement that a broker can no longer cover a client's obligation with the broker's own bond; for the point that a broker cannot create an importer's CARM account and that authority must be delegated by the importer; and for the description of cash security being sized against highest monthly payables. Note: industry commentary from parties with a commercial interest in importers acting. Useful for describing the practical arrangement and not authoritative on any legal point.
- A logistics company blog post citing Customs Notice 25-32 for amendments to section 17 of the Customs Act effective 1 January 2026, to the effect that the importer of record is jointly liable for duties, taxes and post-accounting adjustments and that the entity named on the accounting documents owns the debt. Note: the weakest source in this article by a wide margin. A single secondary source, several rungs below everything else here. We did NOT obtain Customs Notice 25-32 or the amendment, and we flag this in the body as the one item a reader should verify before relying on it.
- No accounting standard was consulted for this article. The principal versus agent analysis is described from general knowledge in terms of control and the usual supporting indicators, being primary responsibility for fulfilment, risk, and discretion in establishing price. Note: recorded here as a reference deliberately, so that the absence is on the list rather than hidden. We did NOT read IFRS 15 or the ASPE equivalent, we cite no paragraph, and every statement we make about the test should be checked by someone with the standard open.