Summary

  • AFRINIC's informational USD presentation shows membership fees of USD 1,754,902 plus grants of USD 151,989, producing USD 1,906,891 of income. After USD 1,767,422 of operating expense, the registry had USD 139,469 before finance and other items; USD 14,906 of interest income less a USD 140,218 exchange loss created a net finance-and-other cost of USD 125,312, leaving an accrual surplus of USD 14,157.
  • The audited MUR statement reaches the same final result, MUR 424,718, but classifies the MUR 4,206,580 exchange loss inside expenses. Subtracting that loss from audited expenses reconciles the apparently different presentations exactly. This is a presentation bridge, not evidence of an accounting error.
  • The published record identifies currency exposure and says the Board reviewed and agreed financial-risk policies, but it does not provide the relevant policy, exposure schedule, approval chain, limits, hedging or no-hedging rationale, realised-versus-unrealised split, or cash-settlement bridge. Those absences prevent a judgment about control effectiveness or wrongdoing; they do not prevent a precise demand for accountability.

Four reported numbers contain the central fact of AFRINIC's 2009 financial year. Membership fees of USD 1,754,902 and grants of USD 151,989 added to total income of USD 1,906,891. Total operating expenses of USD 1,767,422 left USD 139,469 before finance and other items. Interest income of USD 14,906, set against a foreign-exchange loss of USD 140,218, produced net finance and other costs of USD 125,312. Deduct that USD 125,312 from USD 139,469 and the remaining surplus is USD 14,157.

The sequence matters more than the last number alone. Read only the final surplus and AFRINIC looks as though its ordinary activity almost broke even. Read the classification bridge and a different picture appears. The service operation produced a positive result. A currency-risk channel then consumed 89.85 percent of that operating result. The exchange loss itself was slightly larger than the operating result, at 100.54 percent, while interest offset 10.63 percent of the loss. The final surplus represented about 0.74 percent of total income.

These percentages describe the scale of an accrual result; none describes cash generated, cash spent, or money available for distribution.

That separation is the article's subject. It is not a complaint that AFRINIC made too little money, nor an invitation to treat a private registry as a public revenue authority. It is an inquiry into how a member-funded bookkeeper distinguished the economics of its registry service from the financial exposure created by the currencies in which it billed, held balances, and reported. The published accounts make the first distinction possible. They do not carry the reader through the policy, approval, monitoring, and settlement evidence that would make the second part assessable.

Two presentations, one final result. The documentary layers must be kept apart. Ernst & Young's auditors' report covers AFRINIC's 2009 financial statements in Mauritian rupees, then AFRINIC's functional and presentation currency. The report is dated 28 June 2010 and identifies the audited statements as pages 5 through 21. It says the statements were prepared under International Financial Reporting Standards and the Mauritius Companies Act 2001, and gives an unqualified opinion that they presented a true and fair view of financial position, performance, and cash flows for the year ended 31 December 2009.

A separate one-page summary converts and rearranges the audited figures in both MUR and USD. AFRINIC's finance index describes a USD summary of this kind as an informational presentation without official status. The 2012 annual report later republishes the 2009 USD figures in a multi-year table. That later republication is evidence for the comparative numbers; it is not the audit opinion on 2009. The audit opinion belongs to the contemporaneous MUR statements.

The distinction is not pedantic. It controls what each line can prove. The USD presentation is a useful analytical map because it places the exchange loss outside operating expense. It does not become AFRINIC's functional-currency accounts merely because its columns are easier to read internationally. Conversely, the audited MUR face is authoritative for the financial statements but less revealing about the operational-versus-currency split until its expense line is reconciled to the summary.

On the face of the audited MUR comprehensive-income statement, income was MUR 57,206,732. Administrative expense was MUR 34,795,197 and distribution expense was MUR 22,434,049. Together they were MUR 57,229,246, leaving a deficit of MUR 22,514 before finance revenue. Finance revenue of MUR 447,232 then produced the final surplus of MUR 424,718. That path can look inconsistent with the USD summary's positive result before finance and other items. It is not.

The summary reports reclassified operating expenses of MUR 53,022,666. Subtract that figure from income of MUR 57,206,732 and the operating result is MUR 4,184,066. It separately displays an exchange loss of MUR 4,206,580. Finance revenue of MUR 447,232 less that exchange loss equals a net finance-and-other cost of MUR 3,759,348. MUR 4,184,066 less MUR 3,759,348 produces the same final MUR 424,718 surplus.

There is an exact bridge between the two layouts. Audited expenses of MUR 57,229,246 less reclassified operating expenses of MUR 53,022,666 equal MUR 4,206,580: precisely the exchange loss separated in the summary. Once the classification is exposed, the apparent contradiction disappears. The audited face embeds foreign exchange in administrative and distribution expenses, creating a MUR 22,514 pre-finance deficit. The summary extracts that item, revealing a MUR 4,184,066 result from service operations before finance and other items. Both presentations end at MUR 424,718.

This identity is evidence of presentation reclassification. It is not evidence that either presentation contains an accounting error. It is also not permission to call the audited face's pre-finance line a positive operating surplus without explaining what has been moved. The useful conclusion is narrower: removing the exchange loss from operating expense allows the reader to see that the registry-service operation and the currency-risk outcome were economically different channels.

The conversion is a display, not a transaction history. The summary uses an apparent conversion of 30 MUR to one USD, subject to normal whole-dollar rounding. Thus MUR 4,184,066 appears as USD 139,469, MUR 4,206,580 appears as USD 140,218, and MUR 424,718 appears as USD 14,157. That display rate should not be mistaken for the rate applied to a particular invoice, bank balance, settlement, or remeasurement. No transaction-specific exchange rate can be recovered from a presentation conversion.

Currencies therefore need to remain attached to every important figure. USD 140,218 is the converted summary's exchange loss; MUR 4,206,580 is the amount identified in the MUR reconciliation. USD 14,157 is the converted final surplus; MUR 424,718 is the audited final surplus. Treating the numbers as interchangeable without identifying the presentation would erase the distinction between the formal accounts and the informational summary.

One small discrepancy reinforces that caution. The summary presents year-end cash as MUR 38,119,090 and USD 1,270,636, while the audited cash figure is MUR 38,119,091. The difference is one Mauritian rupee. It should be preserved as a display difference, not silently corrected. Financial records often contain rounding or conversion seams that are immaterial to the conclusion but material to honest reconstruction.

The 2008 comparator carries the same lesson. The contemporaneous summary reports 2008 total income of USD 1,429,862 and operating expense of USD 1,311,706, leaving USD 118,156 before other items. Interest income of USD 39,270 plus an exchange gain of USD 135,001 made USD 174,271 of other income and a final surplus of USD 292,427. In the later 2012 roll-up, 2008 operating expense appears as USD 1,311,707 and the surplus as USD 292,426. Each differs by one dollar from the contemporaneous display.

Those one-dollar differences are later comparative rounding or presentation differences. They are not a basis for harmonising the documents by editorial preference. More importantly, 2008 offers a bounded directional comparison: currency effects helped rather than hurt that year's displayed result. It does not supply a long financial history, and it does not explain why the 2009 loss arose. An exchange gain in one year and a loss in the next can be consistent with ordinary exposure to a changing currency. The comparator establishes variability, not culpability.

An accrual surplus is not cash. The most dangerous shortcut would be to describe the USD 14,157 as cash generated or the USD 140,218 as cash paid. Neither statement follows from these accounts. A surplus is an accrual measure of performance. Foreign-exchange accounting can include remeasurement of monetary balances at a reporting date. A reported exchange loss can therefore arise without an equal cash payment during the year. The published record does not give the realised-versus-unrealised split, the transaction dates, the settlement dates, the counterparties, or the cash flows tied to the loss.

The cash-flow statement tells a separate story. It reports MUR 3,601,724 of net cash used in operating activities and MUR 30,821,573 of net cash from investing activities. The investing inflow included a MUR 31,343,218 deposit withdrawal. The two broad cash-flow components reconcile to a net increase in cash of MUR 27,219,849: negative MUR 3,601,724 plus positive MUR 30,821,573. With opening cash of MUR 10,899,242, the audited ending cash was MUR 38,119,091.

That movement must not be merged with the MUR 424,718 surplus. Operating cash use is not the same thing as an operating loss; investing cash generated by withdrawing a deposit is not income; an ending cash balance is not the year's surplus; and a foreign-exchange loss is not automatically a cash settlement. The accounts give each measure a different job. Governance analysis should do the same.

Nor does the positive ending cash position make the small surplus irrelevant. AFRINIC reported total equity of MUR 43,122,532 and year-end cash of MUR 38,119,091. Those figures make insolvency an unsupported description of the USD 14,157 result. They do not answer how the currency position arose, which authority governed it, or whether the risk was being monitored against explicit limits. Solvency and treasury accountability are different questions.

What the notes say about exposure. The audited notes describe an international operation that invoiced a significant number of customers in US dollars while keeping MUR as its functional and presentation currency. They identify exposure primarily to the US dollar and state that an appreciation of MUR against USD could reduce revenue. This gives the exchange result a plausible structural setting. It does not reveal the relevant currency balances, invoice population, bank holdings, maturity profile, or transaction sequence.

The notes also provide a sensitivity exercise. A 10 percent movement in the USD exchange rate was shown as affecting profit before tax by MUR 4,243,542 in either direction, with other variables held constant. A 10 percent EUR movement was shown as affecting profit before tax by MUR 127,211. The USD sensitivity is close in scale to the reported MUR 4,206,580 exchange loss, but proximity is not identity. The sensitivity is a modelled counterfactual. It is not a prediction, a realised loss, a hedge outcome, a cash payment, or an explanation of the recorded loss.

That distinction prevents an alluring but unsupported inference. One cannot work backwards from the MUR 4,243,542 sensitivity to invent the balances that generated the MUR 4,206,580 loss. The accounts do not disclose which receivables, deposits, payables, or other monetary items were exposed, nor how long they remained exposed. The sensitivity says the institution was materially responsive to exchange-rate movement. It does not identify the actual mechanics of 2009.

The notes say the Board reviewed and agreed policies for liquidity, foreign-currency, and credit risks. This is a recorded institutional statement and should be credited as such. It establishes that the accounts represented financial-risk policy as a Board responsibility. Yet it gives no policy text, meeting, approval date, currency limit, investment limit, delegation, exception process, monitoring interval, or compliance report. The statement is the beginning of an authority trail, not the whole trail.

The financial statements were approved by the Board on 28 June 2010 and carry two director signatures. A note says they were authorised for issue by a directors' resolution on that date. Those facts prove the recorded approval of the statements. They do not prove that the Board specifically approved every underlying currency position, that a committee monitored it, or that all intended controls were operating. Approval of accounts after year-end is not a substitute for contemporaneous treasury authority during the year.

A plan records intention, not completion. AFRINIC's 2009-2011 strategy plan is useful because it shows the organisation knew financial architecture required explicit work. The plan assigned finance and management tasks concerning accounting documentation, financial reporting, internal control, a treasury policy for investing surplus cash, reserve accumulation toward a Board objective, and a risk-management document for Board consideration.

Those assignments are evidence of intended control work. They are not evidence that the work was completed, adopted, implemented, monitored, or tested. The plan even contains the impossible printed date 31/11/2009. That source defect must remain visible rather than being silently repaired to a plausible date. A defective date does not nullify the plan; it does underline why a plan and an execution record are different kinds of evidence.

The distinction has practical consequences. A treasury-policy task might result in an approved policy, a rejected proposal, a delayed paper, or no completed document. A reserve objective might guide decisions without becoming a binding limit. A risk-management document prepared for Board consideration might be adopted, amended, deferred, or ignored. The published sources do not tell the reader which of those paths occurred. Treating an assigned task as a completed control would manufacture the evidence under review.

The audit opinion does not fill the gap. The auditors explicitly said that their consideration of internal control was for designing audit procedures, not for expressing an opinion on the effectiveness of internal control. An unqualified audit opinion supports the fair presentation of the financial statements under the stated reporting framework. It is not a general certificate of treasury prudence, control effectiveness, governance quality, or legitimacy. Nor does it validate the lawful approval of every transaction beyond the signatures and resolution statements actually recorded.

This is an important boundary because official documents have two opposite risks. A hostile reader may dismiss them because they were produced by the institution under scrutiny. That would discard valuable evidence. A deferential reader may treat them as self-proving every institutional claim. That would grant them too much. AFRINIC's documents prove the accounts, disclosures, signatures, descriptions, and acts they record. They do not prove an unseen approval chain, the validity of every underlying transaction, the wisdom of every exposure, fairness, necessity, public mandate, or good governance.

The strongest defence deserves full weight. Currency exposure was structurally plausible, not inherently suspicious. AFRINIC billed a significant number of customers in USD while its functional accounts were in MUR. Any organisation with that mismatch may record gains or losses as exchange rates move. A reported loss can be driven by routine remeasurement; it can reverse in a later period; and it need not indicate speculation, a failed hedge, or a cash outflow.

Management may also have prioritised liquidity and operational continuity over eliminating exchange variation. Hedging can carry cost, complexity, counterparty exposure, and its own liquidity demands. The published evidence does not establish what instruments were available to AFRINIC, on what terms, or whether a natural hedge existed. It would therefore be wrong to say the loss was avoidable, unhedged, negligent, unauthorised, or caused by a breach.

The institution disclosed the currency risk, a quantitative sensitivity, the exchange loss, the Board's stated responsibility for risk policies, the final cash position, and the equity position. Its accounts were audited, approved by the Board, and unusually detailed for reconstructing the classification. The final result was positive. Nothing in the published figures proves insolvency, service failure, illegality, corruption, fraud, bad faith, concealment, waste, or misconduct.

Detailed mandates may also contain legitimate confidential information. Publishing bank account identifiers, member invoice records, individual counterparties, or commercially sensitive terms could create security and privacy risks. Accountability does not require indiscriminate exposure. Aggregate currency balances, authority lines, approved limits, exception records, and performance against policy could answer the governance question without exposing member or bank data.

This defence is strong precisely because it separates an ordinary financial risk from an extraordinary allegation. But it does not answer the evidence demand. A natural exposure explains why a loss was possible; it does not show how that exposure was authorised, measured, bounded, or reviewed. An audit opinion on the statements explains why the figures deserve weight; it does not supply an opinion on control effectiveness. Positive cash and equity show resilience; they do not convert a missing treasury trail into a documented one.

The missing bridge. The first absent item is a currency-by-currency balance schedule tied to the reported loss. Without it, the reader cannot identify the aggregate exposures or understand how billing, receivables, deposits, liabilities, and settlements interacted. The second is a realised-versus-unrealised reconciliation. Without that split, the exchange loss cannot be mapped responsibly onto cash.

The third gap is decision authority: the Board-approved treasury policy, the resolution adopting it, any committee mandate, delegated limits, named roles, exceptions, and compliance reporting. The fourth is the hedge decision: an actual hedge, forward contract, derivative, natural-hedge calculation, or documented explanation for choosing not to hedge. The fifth is temporal monitoring: monthly or quarterly variance reports that distinguish service operations, finance effects, and cash settlement.

The sixth gap is transaction and settlement evidence. The published record does not identify dates, counterparties, the invoice population, the balances remeasured, or cash paid and received in relation to the loss. The seventh is independent control testing. The audit did not offer an opinion on control effectiveness, and the strategy plan does not prove completion of its intended work.

Taken together, these absences explain both what can and cannot be concluded. The record supports a positive reclassified service-operating result, a large separately displayed exchange loss, a small final accrual surplus, and a Board-level financial-risk responsibility statement. It does not support a verdict about whether the loss was realised, avoidable, authorised under a particular policy, hedged, improperly monitored, or caused by a control failure.

Why classification changes the governance question. Suppose the reader looks only at the final USD 14,157 surplus. The apparent problem is an institution whose entire year of activity barely covered its costs. That framing would naturally direct attention toward membership pricing, grants, staffing, or programme expense. The reclassified presentation changes the diagnosis. It shows USD 139,469 before finance and other items, so the immediate question becomes how a positive service result was exposed to a net USD 125,312 finance-and-other cost. The policy response should then begin with currency and treasury evidence, not with an unsupported claim that registry services were inherently uneconomic.

The distinction could affect member decisions. A fee rise aimed at repairing an alleged operating shortfall would answer the wrong question if the service operation was already positive. An arbitrary cut to core registry work could likewise weaken the function without reducing the currency exposure. Conversely, a treasury policy designed without regard to billing could miss the source of the mismatch, because the exposure arose in an institution whose customer invoicing and functional currency were different. Service economics and treasury risk are analytically separate, but they still meet on the same balance sheet.

That is why the classification bridge is more than a presentational curiosity. It allocates the problem to the decision layer best placed to explain it. Membership fees, grants, and operating expense describe what it cost to run the service and what the service brought in. Interest, exchange effects, deposits, currency balances, and settlement choices describe how financial assets and liabilities interacted with exchange rates. Board and management responsibility may span both layers, yet review becomes sharper when each layer has its own evidence.

A complete reconstruction would arrange the evidence in three columns. The first would contain recorded facts: MUR as functional and presentation currency; substantial USD invoicing; the MUR 4,206,580 exchange loss; MUR 447,232 of finance revenue; the MUR 424,718 final surplus; Board approval of the statements; and the statement that the Board reviewed and agreed financial-risk policies. The second would contain derived identities: the MUR 4,206,580 classification bridge, the MUR 4,184,066 reclassified operating result, the negative MUR 3,759,348 finance-and-other result, and their reconciliation to the final surplus.

The third would contain evidence not published: balances, transactions, settlements, policy text, approvals, limits, exceptions, hedging decisions, monitoring reports, and control testing.

Keeping those columns separate prevents two opposite errors. One is to push derived arithmetic beyond its evidential capacity, treating a reconciliation as proof of negligence or an unauthorised position. The other is to pretend that missing control evidence makes the reported numbers unknowable. The numbers are knowable. Their classification is reconcilable. What remains unknown is the decision chain behind the exposure and the cash chain behind the accounting loss.

The same method resolves the relationship between sensitivity and outcome. The MUR 4,243,542 USD sensitivity belongs in the recorded-fact column as a modelled effect of a 10 percent movement with other variables held constant. Its closeness to the MUR 4,206,580 loss may guide a reviewer toward the currency note, but it does not migrate into the derived-identity column: there is no arithmetic showing that the sensitivity generated the loss. Nor can it fill the unpublished-evidence column. Only underlying balances and movements could do that.

This three-part ledger also gives a fair test for any institutional answer. A response that repeats the audited final figure contributes no new control evidence. A response that points to Board approval of the financial statements proves the recorded approval, but not the missing contemporaneous mandate. A response that produces an actual dated policy, aggregate exposure report, limit record, and realised-versus-unrealised bridge would close specific gaps. Accountability advances when new evidence moves an item from the unknown column into the recorded one, not when rhetoric is added around an old total.

The appropriate answer is therefore an evidence request, not an accusation. AFRINIC could preserve and republish the 2009 source documents; publish the treasury policy and approval history if they exist; provide aggregate currency exposures and authorised limits; reconcile realised and unrealised effects; show cash and deposit movements separately from surplus; identify Board and committee mandates; disclose exception approvals; and commission independent testing of whether the controls described were operating.

That evidence set would permit a member to trace a consequential financial outcome through actor, mandate, exposure, decision, monitoring, accounting result, cash effect, and remedy. It would also permit correction if a policy was absent, unclear, or poorly followed. If the records no longer exist, that fact should be stated. An unfilled gap should remain an acknowledged gap rather than being covered with retrospective confidence.

Why a registry's modest legal character matters. AFRINIC is a private bookkeeper, coordinator, and contractual service provider. It records number-resource claims, invoices members, holds funds, and provides registry services. Those functions do not create sovereignty, public jurisdiction, legislative competence, regulatory power, police authority, prosecutorial authority, judicial power, punitive power, confiscatory power, or property-title authority.

The same restraint applies to finance. A Board may govern the private corporation within its lawful corporate and contractual authority. It does not thereby speak for a continent, acquire public mandate, or become an owner of the networks and resources recorded in its systems. A Board-approved budget or financial statement is a corporate act. Its legitimacy and scope come from traceable private-law authority, not from ritual invocations of a regional community.

This is why financial accountability must not be turned into registry punishment. A question about exchange exposure does not justify excluding members, freezing registry accounts, revoking number resources, seizing assets, confiscating property, or degrading live network records. AFRINIC is not a court or enforcement agency. If law has been broken, lawful authorities and independent adjudication provide the forum. The registry's own role remains narrow: preserve an accurate ledger, execute bounded contractual services, protect continuity, and correct its own records and controls through evidence.

The doctrinal point is not decoration around the arithmetic. It is what makes the arithmetic governable. A private administrator handling member-funded cash needs authority-liability symmetry: a traceable mandate, bounded discretion, reliable evidence, monitoring, correction, and a remedy proportionate to its actual role. The more consequential the financial or operational decision, the less acceptable it is to rely on atmospheric claims of stewardship or community approval.

Heng Lu's work supplies this controlling boundary: a ledger is a service to reality, not the source of reality, and a bookkeeper does not become sovereign through proximity to valuable records. NRS supplies a first-class demand for transaction-level authority evidence. BTW supplies the connection between control and liability. LARUS supplies the operator-continuity reason that financial and governance decisions at a registry must remain precise, reviewable, and non-destructive. These roles reinforce the accountability test rather than displace primary financial evidence.

NRS advocates, researches, convenes, and represents expressly authorised members. It does not operate a number registry, RPKI, WHOIS or RDAP, appeals, settlements, elections, custody, or continuity services. Its accountability demands therefore cannot be treated as substitute operational records. They identify the questions a private registry should answer; they do not manufacture the underlying 2009 evidence.

None of the available Heng Lu, NRS, LARUS, or BTW materials independently proves the exact 2009 currency transaction, hedge decision, treasury policy, Board approval, policy compliance, or cash settlement. That event-specific gap must be stated without demoting those sources to hostile footnotes. Their institutional analysis controls the boundary of authority and remedy; the 2009 AFRINIC accounts control the numerical reconstruction. The result is a disciplined division of labour between fact and doctrine.

The 2009 story can now be stated without inflation. AFRINIC's service activity generated USD 139,469 before finance and other items in the informational presentation. A net finance-and-other cost of USD 125,312, dominated by a USD 140,218 exchange loss and partly offset by USD 14,906 of interest, reduced that result to an accrual surplus of USD 14,157. The audited MUR presentation reaches the same MUR 424,718 result through a different classification, and the MUR 4,206,580 bridge reconciles the difference exactly.

The arithmetic tells us where the result changed. It does not tell us who authorised the exposure, under which policy, within what limits, with what hedge decision, or with what cash consequence. That is not a reason to invent a scandal. It is the reason to demand the missing ledger of authority.