Joi Ito's Web

Joi Ito's conversation with the living web.

Sony and Docomo have announced that they are working together to put contactless IC chips in phones. Sony's FeliCa (type C contactless IC chip) is slowly becoming a defacto standard in Japan. (The government is backing a different standard, type B.) Currently the Japan Railways, AM/PM and others are using it for payments. Many companies use it for company ID's. The problem is that you can't see how much is left in your card and it's a pain to "charge" the card with more money. Putting it on a phone lets you download money from your bank and see how much is left. I worry about the privacy and security issues, but connecting an RF payment system with a phone totally makes sense.

I have a theory that Docomo has to become an identity/payment company and dump the voice and other bit-pushing businesses and go flat rate or free on the network. Docomo should buy a credit card company and use the bit-pushing business as a stick when collecting money. There are some regulations regarding payment businesses that make it difficult, but I'm sure the government would waive this if there was enough of a social need. Right now, the transaction business that credit card companies do doesn't make money. This has driven credit card companies to become loan companies that lobby the government to allow them to charge crazy interest rates. These interest rates cause people to end up in debt hell and commit suicide. If Docomo replaced credit cards as the primary non-cash transaction, credit system and could use network service termination to lower the collection costs, I bet they could make enough money on the transaction business to cover the bit-pushing.

Docomo is Japan's biggest mobile carrier that does about $8B / yr in data revenues.


We have that in Korea, it's called Moneta, I think. You can even withdraw money from ATMs.

Interesting. Is it a contactless IC chip?

Yeah. It's actually available in mobile phones (SK Telecom) and as a "normal" credit card.
See for an explanation (in Korean...) with a lot of pictures and drawings that help comprehension :-)

Joi, are you still flogging the dead horse of e-cash? This is never going to catch on. The vast majority of Japanese use and will continue to use cash. Credit card use is restricted to deadbeat overspenders and housewives who want the 5% discount for using their Saison cards on every other Tuesday. Docomo, as an NTT offshoot, is basically a quasi-government agency in its corporate culture, and they love these snazzy technology demonstration projects, but like that other quasi-government agency, NHK, and their high-def TV, it will go nowhere.

Before they debuted this technology for the JR train passes, it was launched in Hong Kong as the "Octopus Card."

I'm a skeptic when it comes to mass changes of human behavior but 9 million users of Octopus Cards as of June 2003 is nothing to sneeze at.

Octopus cards are like Suica cards in Japan. Suica is not perceived by users as a substitute for cash, but rather as a substitute for IO cards: prepaid train passes that allow you to skip the ticket machines. Suica allows you leave the the card in your wallet at the wicket (until they go "funny" after a year or two and you have to slap them repeatedly down on the wicket).

Although in theory you can use Suica cards to pay for things in shops, this is not going to happen. People are not going to go to the bank, withdraw cash, go to the Suica machine and pump all their cash into the Suica, and use it to buy stuff. It's just a glorified train pass, a continuation of the Orange Card --> IO Card scam. With Orange Cards, at least you got a discount on your prepaid money. With IO Cards you only get face value. With Suica, you pay them 500 yen for the privilege of using the thing. As with telelphone cards, a nice percentage of the credit never gets used since people lose them.

I think the idea is that it can be a Suica card, but you can view the balance and download money from your bank or credit card into the Suica instead of going to a machine. That seems pretty straight forward. I agree that "digital cash" is a bit of a holy grail, but I think that the Edy card and the Suica are starting to penetrate the market. Hooking them up to a phone makes sense to me.

I'm with Joi. A big thing. Contactless payment, RFIDs, barcode reading. Used to paying our telephone companies a bunch of cash each month, the ease of use, familiarity and trust is there that was perhaps missing with many ecash/similar systems.

But won't just be DoCoMo. Once Vodafone and KDDI introduce flat rate billing and fight harder for number portability, we might get 3 banks/id brokers.

Sounds like the banking equivalent of Japan's N-System that tracks your car all over the country with unmanned cameras and automated license plate recognition. Only now they'll know every train you rode, every Pocari Sweat you bought, when and where. The tax authorities will love it. So will your wife--just being able to peek at your cell phone call history isn't incriminating enough, I guess.

Joshua, I'm also concerned about privacy and security, but at least it will be multi-vendor and you will be able to have several chips if you like. The government is pushing the national ID card to become your payment system. That's much scarier.

I think the DoCoMo/Sony Felica deal is going to be a very important step in the development of e-cash.

Sure people don't think of it as e-cash when they use it on JR but isn't that what makes it a perfect stealth way to get the cards into millions of people's wallets?

Having them in every cellphone will be a great way to get them out as well -- and that's always been the hurdle e-cash systems face along with retailer acceptance.

Everyone who uses Suica is sold on the convenience and I am sure people will be persuaded to use it for cash payments when they start seeing it in station Kiosks.

Having the chip in the phone also makes a lot of sense for new applications. You can just download the relevant app into the Felica chip -- a standard Felica chip has space for 40 applications -- and you're set to go.

Theoretically/technically you could do things like download a JR-West app into your phone before making that week-long business trip to Osaka, add your corporate security pass app to the phone and do away with the ID card or even add in an Octopus app before your trip to Hong Kong.

With the network link for topping up the e-cash cards it makes it into a very powerful technology and one that I think will prove a huge success.

Also remember e-cash systems like Edy are not trying to replace credit cards but are aiming for payments of around 3,000 yen or less. They are targeting pocket change type transactions.

The credit card companies, and DoCoMo, KDDI and Toyota are already shareholders in Edy.

It shouldn't be forgotten either that the DoCoMo/Sony company, Felica Networks, will be developing the mobile technology but it won't be restricted to DoCoMo. The other carriers will get a chance to license this.

As for privacy and security, the cards themselves are anonymous. JR or Edy (e-cash) have no idea who has which card and there is nothing to stop you using several or regularly changing them.

Purchases are not even reconciled, in the case of Edy, on a real time basis but checked at the end of each day. I do not believe there is any record, at Edy, of the products purchased.

When you talk about putting them in a cellphone then it becomes possible for the cellphone company to talk to Edy and match up records and then go to the convenience store and find out what was purchased ... but it doesn't seem very likely.

And anyway, if you have such distrust of your carrier you probably don't own a cellphone anyway because you don't want your progress around the city tracked.

There are some concerns about what happens if you lose your phone but at least the carrier can disable these apps and the phone over the network. Its a lot better than losing a wallet full of cash and credit cards but we all manage to overcome that worry and carry our wallets all the time.

That's my 10 cents worth, anyway. Wow, did I really write that much ;)

The entities currently pushing e-wallet projects all over the world, be it in Asia, Europe or the US still don't get it. All too often, the business case is geared towards the convenience, cost savings or profit accruing to the payment infrastructure suppliers.

For banks and retailers, cash management costs -- e.g. armored vehicles -- and capital opportunity costs -- banknotes and coins in cash registers and ATMs don't bring any interest revenu -- are significant.

E-money, OTOH, can be pooled and credited to an interest-bearing account almost instantaneously, making cash management mush more efficient. There would be less need for expensive armed security guards, safes, special transport vehicles etc.

I don't see *why* the consumer should thus *pay* e.g. a yearly subscription so that the banks and the retailers can lower *their* operating costs, especially as e-wallets are of little tangible benefit to the end-user -- i.e. you and me -- today.

There is a conceptually simple way to make e-wallets more attractive than banknotes and coinage, at least for honest people who have no immediate economic incentive to transact ia an underground/black economy: it is to make the e-wallet cash recoverable in case the wallet is lost. Today, losing a wallet containing currency usually means kissing your money good-bye. With a recoverable e-wallet, it should conceptually be possible to invalidate a lost e-wallet and re-credit the money it contained to your bank account. That would be one tangible, and immediately understandable benefit of e-wallets that should be "sellable" to the end-user. E-wallet smartcards should be issued accompanied with a "master" card. That master card usually stays at your home, and you'd just use it to authenticate yourself at an ATM to invalidate and recover any money that's currently left from your lost e-wallet.

Such a scheme obviously requires that every e-wallet transaction be recorded in a central database, so that the remaining amount of every e-wallet is exactly tracked. A telecom link must thus be initiated by the payment terminal to the central database every time an e-wallet is used, which would of course require staggering amounts of telecom bandwidth and processing power. One should note, however, that bandwidth -- even of the wireless kind -- is pretty plentiful, and computer processing power costs keep getting lower and lower. The marginal cost of exchanging a few data packets with a payment terminal can thus be brought to close to zero. Besides, the number of phone calls I give in a day is usually larger than the number of times I make a cash payment. The aggregate computation power required to route each phone call taking place in a country on a given day is staggering, as is the volume of billing data generated for each call, but last time I checked, the phone companies' switches and IT systems were able to keep up with such a demand.

Concerns about privacy can be addressed by recording just the current amount stored in a card -- i.e. not its transaction history -- in the central database, and by making the database operator financed by banks and retailers, but independent of them. Banks would just be authorized to issue new e-wallet IDs, cancel existing ones, top-up the stored amounts and settle their end-of-day net positions vis-a-vis the other banks. Privacy is maintained by not communicating to the central database the actual bank account numbers associated with an e-wallet. There would thus be no means for rogue employees of the central database operator to determine the payment history or owner name associated with a particular e-wallet.

Money laundering concerns can be alleviated by limiting the maximum amount that can be stored in an e-wallet to e.g. US$500, and making impossible the direct crediting of money from one e-wallet to another. Payments between two e-wallets would have to transit through the associated bank accounts first, the withdrawal from one e-wallet and crediting of the associated bank account taking place one day, and the transfer from th payer's back account to the payee's bank account, and subsequent crediting of the e-wallet on the next day. Enforcing a day of delay for the completion of inter-wallet payments would thus add a "friction" to the payment network that would add but negligible inconvenience to honest users, but put a brake on, and hobble the presumably large-scale transactions required with money laundering schemes using e-wallets.

What I would like to see is a payment method combining the privacy protection of e-gold with the universal acceptance of credit cards.