Felix Salmon

The unintended consequences of cheaper remittances

By Felix Salmon
March 5, 2014

Once upon a time, remittances, especially to Mexico, were the next big thing. In 2002, for instance, Bank of America bought 24.9% of one of Mexico’s big three banks, Serfín, mainly for the remittance business:

Bank of America says it will compete with Citigroup for Mexican and Mexican-American customers in the United States. It particularly hopes to win a larger share of fees from the $10 billion in remittances they send to Mexico each year…

Kenneth D. Lewis, Bank of America’s chairman and chief executive, made it clear today that a primary goal of the deal was to gain more Mexican-American customers in the United States and ”dramatically increase our market share” of their remittances.

BofA’s purchase was in large part a defensive move — a reaction to the 2001 acquisition of Mexico’s largest bank, Banamex, by Citigroup. That deal, too, had a significant remittances component to it, and by 2004 Citi and Banamex had launched their Access Account, a product allowing Mexicans to easily send money from any Citibank branch in the US to any Banamex branch in Mexico.

The Access Account competed against a consortium of Mexican savings banks, who had teamed up with US Bank to create a similar product in 2003. But that wasn’t the end of the dealmaking: later in 2004, the third big Mexican bank, BBVA Bancomer, bought Texas bank Laredo National Bancshares for $850 million, with the aim of increasing Bancomer’s 40% share of the remittance business.

The banks were racing into the remittance market because it held a huge amount of promise: it was growing fast, the incumbents (Western Union and MoneyGram) were easy to undercut, and the potential profits were huge: after all, to a big bank, the marginal cost of moving money from the US to Mexico is essentially zero.

The promise of cheaper, next-generation remittances was so great that the World Bank, in 2009, set what it called a “5X5 Objective”: that it would reduce the cost of remittances by five percentage points in five years. By 2014, remittances would cost only 5%, rather than the 10% prevailing in 2009.

The objective was entirely achievable — and indeed, last year, looking at the depressed MoneyGram share price, I blithely declared that “money transfer is in the process of being disrupted”.

Certainly the growth in remittances, over the past five years, would more than allow for economies of scale. While international capital flows have fluctuated, remittances have been growing steadily, and have remained above all debt and portfolio equity flows every year since the financial crisis. Here’s the flows chart, from the World Bank:

Screen Shot 2014-03-04 at 7.48.07 PM.png

But here’s the chart showing whether the 5X5 objective is going to be met — and it’s not pretty. The reduction to an average cost of 5% isn’t even close to being achieved.


Screen Shot 2014-03-04 at 7.50.41 PM.png

In Mexico, remittance flows are falling, up even as the number of migrant workers in the US now comfortably exceeds the levels seen before the financial crisis. And the banks, who were once so excited about this market, are packing up shop:

Banamex USA underwent a downsizing last year. The company had a sizable business in taking money from third-party agents in the United States and then remitting the money back to an extensive network of Banamex bank branches in Mexico, industry experts say.

But now Banamex USA will transfer money from the United States to Mexico only from its own customers, a spokeswoman said. Last year, Banamex USA also reduced the number of its branches in California, Arizona and Texas, three states with large Mexican immigrant communities, to three from 11.

Citigroup said the changes at Banamex USA are part of the bank’s global restructuring of branches and businesses. But industry participants suspect that the moves may have more to do with avoiding the costs and risks of trying to meet anti-money-laundering regulations.

The problem with remittances, it turns out, is that such operations have a habit of getting hit by anti money laundering probes. The current problems at Banamex, for instance, come in the wake of similar issues at Western Union, which stopped using thousands of agents in Mexico who couldn’t meet regulatory-reporting requirements. On top of that, because there’s no shortage of smaller companies trying to compete on price, busy corridors like US-Mexico are now actually pretty cheap: the World Bank’s Dilip Ratha told the NYT last year that the cost of a $300 transfer is now only around 2%.

Indeed, if you look at the World Bank report, a curious phenomenon emerges: remittances seem to be growing fastest where they’re most expensive, and falling where they’re relatively cheap.

What this says to me is that if the World Bank wants to maximize remittance flows, maybe concentrating on bringing the price down is not the best way of doing that. Financial services to the poor are nearly always expensive, and the rich tend to have a very understandable and predictable reaction when they see that: they want to bring the price down, for the sake of the poor people. All too often, when that happens, the supply of that service tends to dry up — or the poor continue to pay more than they strictly need to.

Here’s a theory: when the cost of remittances is high, the providers of those remittances have every incentive to make it as easy as possible for as many people as possible to remit as much money as possible back home. And when the cost of remittances falls, those incentives weaken, and it’s easier to sever ties to merchants and generally discourage the use of services which you formally pushed aggressively. Maybe remittance services are sold, just as much as they are simply purchased. As a result, when they’re cheaper, and not sold as hard, the migrants end up spending more money where it’s earned, and sending less back home.

Most of the new companies competing on price against Western Union are doing so with mobile apps and the like: they try to make it as easy to send money home as possible. Maybe it’s as easy as just pressing a few buttons on your phone. But I suspect that what such services are not doing is adding much in the way of behavioral layers: they’re not giving people an incentive to send as much money home today as they possibly can. After all, if it’s that cheap and easy, why not send less today, and then more tomorrow, if it’s still there?

There are still a lot of areas of the world where the cost of remittances is too high, of course. But when it comes to Mexico flows, I don’t think that’s necessarily the problem. Not any more, anyway. Instead, if anything, the cost of remittances might be too low: it doesn’t give banks an incentive to be active participants, given (a) the headaches involved with respect to money-laundering regulations, and (b) the up-front cost of enticing migrants to use the service in the first place. So the banks will keep the product in place — they just won’t make it particularly easy to use, and they won’t promote it aggressively. And they won’t be too upset if usage declines.

Update: Timothy Ogen replies.

6 comments so far | RSS Comments RSS

After reading this, I’m unsure if it’s harder for people in the US to send money back to Mexico, so it’s hard for me to see why lower fees are a problem. Sure, some banks have backed out, but has that really been a problem for remitters?

Also, the correlation between rates and growth might simply be statistical or the result of an inefficient market. Are the transfer locales with the highest growth really small, making high rates of growth easy and also implying high costs? Or are they served by an effective monopoly, meaning a competitor can increase usage with just a slight lowering of prices?

Posted by AngryInCali | Report as abusive

Hi Felix, interesting blog. I was wondering whether you heard about such thing as “the financial crisis”? Is it possible that there is no correlation between prices decreasing and flows dropping in US to Mexico, and that another factor is contributing to the reduction of flows instead?

Posted by marconick | Report as abusive

I think you are missing the cause-effect. As remittances between two countries grow, the cost to send that money goes down. Invariably, remittances for that pair of countries will stop growing, level off, or decline. As new countries emerge as destinations of outgoing funds, their transaction costs will be higher than the traditional destinations. You may also have geographic effects – no matter how electronic business gets, it will be easier to manage transactions between the US and Mexico than the US and China.

Remittances are mostly about to-Mexico transactions. But if you look at the World Bank data: http://blogs.worldbank.org/peoplemove/es timating-outward-remittance-flows-from-t he-us-over-100-billion-a-year
nearly a third of US remittances go to Asia. I would imagine that just comparing costs between Mexico, China, India, the Phillipines and Vietnam, costs to send money to Mexico would be considerably lower than the others.

Posted by winstongator | Report as abusive

What’s 50 + 67 equal? I believe it’s 117 Right? So then the year 1967 plus 50 years would equal the year 2017 right?

And the year is now 2014 so in 3 more years the American Freedom of Information Act allows the world to view the first Photographs that must be made public of the USS Liberty Massacre where Israel ordered its Army, Navy, and Air Force to destroy our American war vessel by bombs, Torpedoes, Strafing, and Shooting the American Officers and Enlisted Men in International Sea Waters that immediately killed 37, and wounded 170 that have died and been dying from their wounds in the 1967 Israeli 3 Day War.

No Enteral Flame has ever been erected in Arlington National Cemetery for these brave dead American Soldiers, no mention of their sacrifices in any American High School History Books, no President, Senator, or Representative has shown any public display of gratitude for their courage in this conflict, all those who were there and died have been all but forgotten but in 3 years the world will come to know the truth.

President Johnson was suspected of the JFK assasination by conspiracy theorists and the conspiracy to keep quite the USS Liberty Massacre for the last 50 years came on President Johnson’s shift so is there a link between the two conspiracies with the same man involved?

Posted by 1justmyopinion | Report as abusive

Why is it a “curious phenomenon” that prices rise as volume goes up? Didn’t the housing market do the exact same thing?

Posted by FDum | Report as abusive

I wonder if banks are going to find themselves bypassed when remitters use bitcoin (et al).

Posted by markhahn | Report as abusive

Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/