One important distinction is that Bitcoin wallets are free where bank accounts aren't. Phones that can display QR codes have become cheap now.
Ignoring fees for a moment, if there are Bitcoin ATMs on both ends of the remittance transaction, it's pretty easy for the participants. Probably even competitive.
The problem with Bitcoin ATMs is the logistics of cash.
A Bitcoin ATM network that operates at a scale large enough to displace or even compete with the remittance networks in Mexico (and other beneficiary countries) would by necessity draw on the banking system to supply it with cash. This is because all of the cash being withdrawn would have as its ultimate (albeit intermediated) source Bitcoin purchases in the US (or other sending countries).
Think of it like this: the Bitcoins being sold in Mexico in exchange for pesos at these ATMs are being purchased by the ATM owner, who will then need to sell them in order to resupply the ATMs with pesos. Who does she sell them to? Without a massive (i.e. hundreds of millions of pesos per day) local demand for Bitcoins in Mexico, her buyers are going to have to be foreign--most likely the the immigrants in the US who are sending money home.
So, she sells those Bitcoins to Mexican immigrants in the US, for dollars. And now she needs to buy pesos, in the form of bills/notes, in exchange for dollars--at which point she is in the exact same situation as any standard remittance provider today!
Everywhere in the world, one of the primary roles of banks is to manage cash logistics, and to act as the retail face of the central bank, which acts as the ultimate sink for excess cash, and as the ultimate source of cash in the event of a shortage. As a result, the banks, and then ultimately the central banks, regulators, and governments generally, act as gatekeepers for any large-scale financial enterprise requiring the disposal and generation of physical cash.
What's more, when you are talking about billions of dollars exiting the US in favor of Mexico, there is only one way that money can move, which is through the central bank, via the banking system.
The only way that Bitcoin can displace the current remittance providers is if a large local demand for Bitcoins already exists in beneficiary countries, such that the inflow of Bitcoin could be absorbed locally, or used by recipients without being cashed in. Either that, or central banks and regulators will have to embrace it--or at minimum tolerate it.
That's not to say it can't happen. It just can't happen without official sanction first, before a local bitcoin market has established itself.
Potential demand for Bitcoin in these countries could be companies that buy from the US and receive a discount for paying with Bitcoin. An intermediary buys Bitcoin for the receiver of remittances and sells to those companies.
That could work. Some guy with a shop in Mexico selling say used iphones could provide a cash sevice giving people pesos in exchange for payment in bitcoin and then use the bitcoins to purchase iphones from the US. Assuming there are sellers in the US who will ship goods to Mexico for bitcoin.
Why would you ignore fees? Fees are the core factor of any payment/remittance business. The possibility of such an 'Bit-ATM' network doesn't depend on any technical issues, but on the amount of fees that can be extracted; only the volume and rate of fees determine if it's realistic or not.
Bank accounts also are generally free or near-free worldwide, it makes sense for every minimum wage earners to have a bank account, around here they do unless they're working illegally. If they're unaffordable in some places, then maybe that's the thing to solve - it doesn't actually cost much money to keep a "account" & some digital records; and the clerk-face-time of making such a record is comparable to the clerk-face-time of current money remittance services.
Decentralized P2P bitcoin transfers are not easily regulated; It would be like regulating other P2P technologies like BitTorrent or encryption software. Of course, if a centralized business is in the middle, you can regulate that, but it's not a necessary piece of the solution.
Such centralized businesses will be swiftly shut down by legislation if bitcoin cannot be regulated.
If shutting down the centralized businesses isn't enough, bitcoin will be outlawed. Identifying users is easy enough due to the nature of the network.
What will the government do with Bitcoin the day a report comes out that American citizens have been killed in a terrorist operation financed by bitcoin?
And yet a little "https" logo smiles at me in the address bar as we speak.
Point being that the government can't and doesn't ban everything that can be used for evil, be it from incompetence or rationality. Bitcoin will probably be heavily monitored though, as it is very suitable for that purpose.
HTTPS was crippled, since Netscape shipped only 40-bit RC4 internationally.[1]
PGP's source code was printed so it would fall under First Amendment protections, since binaries weren't legal for export.[2] (See also DJB's Bernstein v. United States)[3]
OpenBSD/OpenSSH is still based in Canada to avoid being subject to the laws.
Isn't it conceivable that the government would implement a similarly short-sighted (and ineffectual) policy if Bitcoins were seriously used to harm the US? :-P
I'm not sure identifying users of the network is quite as easy as you may think it is. If by users you mean those sending or receiving bitcoin transactions. It'd be an interesting challenge to identify the IP address originating even a single Bitcoin transaction. How can you tell whether a node is the originator or simply a relayer of a transaction? For added privacy, run Bitcoin over Tor.
You're not completely wrong. But without Bitcoin being accepted everywhere for everything, dollars will always be involved, and that's when're the regulation can and will live.
Ignoring fees for a moment, if there are Bitcoin ATMs on both ends of the remittance transaction, it's pretty easy for the participants. Probably even competitive.