Point exchange network risk (JP cross-program conversion economics)

Wiki route

This entry sits under loyalty index as the cross-program exchange page — the mechanism layer beneath the program directory in Japan points and loyalty landscape. It pairs with point liability accounting boundary (which defines how a conversion splits into a liability-transfer event plus a settlement leg) and with JMB vs AMC (because airline mileage is the highest-value sink most exchange paths flow toward). For the regulatory edge — where a freely transferable, cash-like point stops being loyalty marketing and becomes a payments topic — route to funds-transfer vs prepaid boundary and payment-license stack.

TL;DR

A point that can be converted into another operator’s point or into airline mileage is no longer a self-contained loyalty liability. Conversion creates a small inter-operator financial network: each edge carries a consumer-facing exchange ratio, a (usually different and confidential) bilateral settlement rate, a liability transfer, and a set of anti-abuse controls. The network is where most of Japan loyalty’s systemic risk concentrates — non-additive liability accounting, FX-like rate exposure, arbitrage / laundering vectors, and the route by which a “marketing” point drifts toward cash-equivalence. Reading any single operator’s point balance in isolation overstates how contained that liability really is.

Why exchange turns a point into a network

A closed point (earn at Rakuten, spend at Rakuten) is one operator’s deferred-revenue liability and nothing more. The moment an exchange edge opens — Rakuten Point ↔ ANA mileage, Rakuten Point ↔ JAL mileage, dポイント ↔ JAL mileage, Pontaポイント ↔ JAL mileage — three things become true at once:

The same yen of purchasing power can exist in two liability lines. During a conversion window the originating operator may not yet have extinguished its liability while the receiving operator has already created one. Outstanding balances across operators are therefore non-additive; you cannot sum operator-reported point balances to get a “Japan point economy” total.
There are two prices on every edge, not one. The consumer sees an exchange ratio (e.g. Rakuten → JAL at 2 points = 1 mile). The operators settle at a separate, contractually agreed rate that is normally not disclosed. The spread is the receiving operator’s acquisition revenue and the originating operator’s release cost.
Asymmetry is deliberate. Reverse edges are usually priced worse and capped. Publicly, Rakuten’s reverse path (JAL miles → Rakuten Points) runs at roughly 1 mile = 0.8 point at scale, versus 2 points = 1 mile inbound — a one-way valve that pushes value toward the higher-margin mileage sink and discourages round-tripping.

This is the concrete content of bucket 5 (“cross-program exchange”) in the accounting-boundary page: a liability transfer plus a settlement leg, happening together.

The five risk surfaces

Surface	What goes wrong	Who absorbs it
Settlement / counterparty	Net-settlement timing gap between operators; one side’s insolvency mid-window	Both operators; ultimately the consumer holding mid-conversion balance
Rate / FX-like exposure	Operator changes consumer ratio or settlement rate; value of held balance shifts	Holder of the soft-pegged point; arbitrageurs on the other side
Breakage mis-estimation	Transferable points sit longer, so breakage assumptions stretch and revenue timing slips	Issuer’s reported revenue under ASBJ Statement No.29 / IFRS 15
Arbitrage / laundering	Rate gaps, promo stacking, or weak identity binding let value be cycled or cashed out	Operators (fraud loss); the funds-transfer / AML regime if value becomes cash-like
Cash-equivalence drift	A freely exchangeable, redeemable point starts behaving like money	Regulatory boundary — Payment Services Act analysis

Settlement and counterparty risk

Exchange edges settle bilaterally and net, not in real time. Between the moment a consumer converts and the moment the two operators settle cash, there is a window in which the originating operator owes the receiving operator. A large, sudden conversion event (a viral “convert before the rate changes” campaign) concentrates this exposure. The receiving operator carries a fresh liability it must honour on redemption regardless of whether the originating operator’s settlement has cleared.

Rate / FX-like exposure

Because each edge has a soft-pegged consumer ratio, a point held because it can be exchanged carries quasi-currency risk. When an operator revises a ratio — or, as ANA publicly signalled for its Rakuten-point exchange, lengthens the transfer window and introduces per-day redemption caps — the practical value and liquidity of the held balance change. Holders chasing the best exchange path behave like FX traders; operators set ratios and caps the way a currency board manages a peg.

Breakage mis-estimation

Transferability extends effective life. A point that can be parked in a high-value sink (airline mileage, a partner program with longer expiry) gets redeemed later, or shifted rather than abandoned. That stretches the breakage assumption that, under both ASBJ Statement No.29 and IFRS 15, governs when deferred point revenue is recognised. Optimistic breakage on a transferable point recognises revenue too early; the exchange network is exactly where that estimate is hardest to defend, and where the disclosure norms pushed by the Cashless Promotion Council / Payments Japan (comparable redemption-rate, expiry, and outstanding-balance reporting) bite hardest.

Arbitrage and laundering vectors

Rate gaps plus weak identity binding are an arbitrage invitation: cycle value across edges to harvest spreads, stack promotional multipliers across the same conversion, or aggregate many small balances into a cash-out path. This is why operators wrap exchange flows in rate limits, conversion increments and monthly caps, identity binding, and throttling of unusual patterns — controls that read as marketing friction but function as fraud and AML defence. The publicly visible guard rails (minimum 50-point exchanges in fixed increments, monthly conversion ceilings) are the consumer-facing edge of this.

Cash-equivalence drift

A point that is freely transferable, broadly redeemable, and effectively refundable starts to behave like money. At that point the loyalty framing fails and the Payment Services Act analysis takes over — prepaid-instrument or even funds-transfer territory. The boundary is set out in funds-transfer vs prepaid boundary and the payment-license stack. Exchange design is, in part, deliberately kept limited (limited-use points, caps, one-way valves) to stay on the loyalty side of that line. The clearest illustration is that limited-use dポイント (期間・用途限定) cannot be converted to JAL miles at all — the restriction is what keeps the promotional grant from becoming cash-like.

Topology of the JP exchange network

The network is not a uniform mesh. It is a set of soft-pegged edges with airline mileage acting as the dominant high-value sink.

Edge (publicly documented)	Consumer direction	Character
Rakuten Point ↔ ANA mileage	mutual; ~2 pt = 1 mile inbound	longest-standing (mutual exchange since 2004)
Rakuten Point ↔ JAL mileage	mutual; ~2 pt = 1 mile in, ~1 mile = 0.8 pt at scale out	added 2022; explicit asymmetric reverse rate + monthly cap
dポイント ↔ JAL mileage	toward miles; base points only, limited-use excluded	telco point into airline sink
Pontaポイント ↔ JAL mileage	mutual; bilateral net-settlement with JMB	mature; predates several peers operationally

The structural reading: common points (Rakuten, d, Ponta) are the broad on-ramps; airline mileage is the deep sink. Value flows toward mileage because that is where per-unit value and emotional redemption value are highest, which is also why the reverse edges are throttled. For the program-by-program map feeding these edges, see Japan points and loyalty landscape; for the airline-side mechanics of the sink, see JMB vs AMC.

Why this matters for JapanFG / financial analysis

Liability is non-additive. A high point balance on one operator’s IR slide is not a system-wide total, and part of it may be in transit to another balance sheet. Aggregate “Japan point economy” figures (the ~2.8 trillion yen Yano Research market sizing is a market figure, not a summed liability) should never be read as a sum of operator liabilities.
Acquiring a point operator imports its edges. A bank or telco buying into a common point (SMFG via V-Point / CCCMK, NDFG via dポイント, Rakuten FG internally) inherits the settlement relationships, the breakage-estimation problem, and the cash-equivalence boundary — not just a marketing asset.
Exchange is the cash-like pressure valve. The richer the exchange network, the closer the most-liquid points sit to the funds-transfer / prepaid boundary. Operators that have layered card, bank, and securities products on top of a point (PayPay FG, Rakuten FG) have the most to lose if a regulator reclassifies a flagship point as cash-equivalent.

Sources

Rakuten Group press release — Rakuten Points and JAL Miles mutual exchange (2022-05-11).
ANA Mileage Club — Rakuten Point exchange terms (consumer ratio, transfer window, redemption caps).
Rakuten Point Club official guidance — exchange minimums, increments, and monthly caps.
ASBJ Statement No.29, “Accounting Standard for Revenue Recognition” (収益認識に関する会計基準) — breakage and contract-liability framing.
Payments Japan Association / Cashless Promotion Council — code-payment disclosure norms.