Ask HN: Bank 2.0. How would you build a better bank?

14 points by dkokelley ↗ HN
I was thinking about consumer banking while reading this thread: http://news.ycombinator.com/item?id=948460

I came to the conclusion that consumer banks have a lot that are in need of fixing/improving. I posted a comment about how I would build Bank 2.0 if I had the means. I figured that I would post a discussion on the subject to hopefully get a lot of input from a lot of smart people.

My question to you is, how would you build your Bank 2.0? How would you make money? How would you be different from other banks out there? You have some seed money (for new banks, this might be about $20M), and your goal is to start a bank in your town. How would you start, and how would you grow?

I'd like to focus this discussion more on consumer banking (think, checking, saving, and small business accounts).

41 comments

[ 3.8 ms ] story [ 89.4 ms ] thread
Take some time to read about what Vernon Hill did at Commerce Bancorp.

Hill came from a non-banking background, he was an owner of Burger King franchises. As a result though, he really figured out that one of the keys to creating a thriving bank was to make a bank that had great customer service.

Mostly, innovations are used to get cheap deposits. If there is a bank that can innovate and do it correctly, then it can work well. Commerce is a great example of one that did achieve it. Their efficiency ratio was higher than most peers but that’s because they kept their bank open 7 days a week and they made it so you really wanted to visit their branches. They used nice colors, had super friendly tellers (they weren’t stuck behind bullet proof windows or a mini gate), and they didn’t make branches look like a post office or a jail cell.

Their whole idea was putting more money towards non-interest expense but they were able to charge less for deposits since they were more convenient. It definitely worked. If you could invest in safe assets you could earn a wider NIM than peers.

HOWEVER:

You can be the most innovative banker in the world but if you make bad loans you’ll go down.

vernon hill writes regularly at http://www.bankstocks.com/
Just be cautious of Tom Brown (the owner of that site). The guy is basically a huge gambler/perma-bull on banks. A lot of his insights were proven to be wrong during the crisis.
To start, my Bank 2.0 would do a lot to change the way credit/debit cards are authorized. clawrencewenham had some great ideas about this here: http://news.ycombinator.com/item?id=948965

Basically the owner of the card would have final control over charges on it. The bank could automatically call or text you and ask for a code before authorizing a transaction, and the cardholder would have white and blacklists with custom rules about when/how much a particular merchant could charge.

In my bank, I would probably stay away from mortgages and credit cards (I could offer them through a 3rd part supplier, but I wouldn't carry any of the balances). This is to make reserves more manageable. Today's banks have massive departments dedicated to managing reserves and working with overnight and discount loans. By eliminating/outsourcing these functions, I think more could be dedicated to serving the customer. Most of the money for the bank would be made through an agreement with someone like Visa or Mastercard, so that whenever someone charges their card, part of the fee that merchants pay goes back to my bank as the card issuer.

I don't see why you would not offer mortgages.

The point of a bank is to have a healthy net interest margin on the funds that they are borrowing and the funds they are loaning out.

Mortgages are like having a long term contract where that margin is secured. Also, think about it. If you are not making mortgages, you're going to be looking at more consumer-focused later credit stage loans.

Commercial real estate, construction, small business loans. All of these are loans that are more difficult to write and typically carry higher loss rates. When I look for banks to invest in, I am typically looking for banks that write mostly single family mortgages because they are more secure than those lending to local businesses/developments.

Also, I think you are confusing most banks with the dealer broker model - Bear/Lehman/Merrill/Goldman/Morgan Stanley. Those banks had massive amounts of overnight lending. Commercial banks like JPM, Wells, and BoA did not. That's why they are able to thrive and consolidate during the crisis.

A great bank will have sticky deposits -- low yield Savings accounts, checking accounts. Not high yield CDs.

I was thinking about mortgages in terms of a bank in my home town. Being a smaller bank, I'm not sure I would like to be carrying a lot of loans, but I could create them for other's to carry in return for a commission. You are right, I might be confusing banks and brokers.
You would be pretty surprised. The loan books on most tiny banks are mostly comprised of single family mortgages.
Having to authorize every damn transaction (even with whitelisting) by phone would make me want to just give up and use cash most of the time. Can you imagine the user experience?

-You can't dare forget your phone wherever you go.

-You close your tab at the bar and you have to text your authorization code in (drunk) while everyone else is waiting in line behind you (because it's closing time and everyone's closing their tab at the same time).

-The authorization time is 60-120 seconds, long enough for you to authorize on the spot before you get your receipt. Your ISP tries charging your internet bill while you're driving, and because you're a safe and reasonable person who doesn't text while driving, the authorization times out.

-Or maybe the authorization time is like 3 hours, which is long enough for you to just plain not pay your bar tab, restaurant bill, or grocery bill, which means bars, restaurants, and grocery stores won't accept your card.

When it comes to recurring bills, here's an easier idea: on your banking history, right where it says "Verizon" or "Comcast" or "Obfuscated Porn Site Billing" or whatever, you can click a button that says "stop authorizing payments to this merchant". As the bank, if you recognized the merchant you'd send them some sort of automated message that said "Our customer, John Doe, the owner of philwelchbankcard account 51431070970963941, has cancelled future payments to you. Please consider this as notice that Mr. Doe has cancelled any and all subscription to your service. You may contact a philwelchbank representative at 1-888-4-PW-BANK to discuss any outstanding charges or early termination fees." So the merchant would have to either send John Doe a bill in the mail or call the bank and say "John owes us $175 termination fee." And then the bank would call John Doe to authorize any final charges regarding termination fees. (The intended effect is to make it more difficult to charge termination fees. It's important that the bank remain on the customer's side, not the merchant's side.)

You bring up a lot of good points. In my mind I was thinking of text (or email, or phone) confirmation for online transactions and merchant-initiated transactions. Basically anywhere you don't physically present the card. Having the card present is a pretty good indication you intended to make the payment. Online orders are done through payment gateways, or going back a bit, manually keyed at a terminal. Those transactions should be confirmed manually or white-listed before they are authorized.
With a long enough confirmation period that would probably work.
Some Korean banks already do this - My girlfriend gets a call from her dad every time she uses his credit card number online, as he gets a text message notifying him it has been used for card-not-present transaction :)

He doesn't need to authorize it, but getting a notification that it has happened at all is better than nothing.

Semi-related: Does anyone know why credit card transactions can be authorized in seconds but ACHs (or whatever underlies online bill payment) take days to go through?
ACH uses the check clearing system (since that's exactly what they are: checks). This system was originally designed in a time when checks needed to physically move from bank to bank, and while there have been attempts to speed it up, debit cards have removed much of the pressure for making this timely.
Check21 should have alleviated this but I think ACH and Check21 are separate rules systems. A bank I use (USAA) lets me scan checks and upload the image to deposit the check. Funds are available instantly (I do not know if they are lending me the funds interest free during the window, or if it is actually cleared instantly).
this is very much dependent on which country you live in. I moved from South Africa to the UK and was very surprised by how long online transfers take to clear. Back home they can clear instantly most of the time, even across different banks. The max time for any online transfer I ever experienced was 24 hours. Here in the UK it seems the minimum is 48 hours, really not sure why.
It probably depends on who is accepting the fraud risk, and how much risk they are willing to accept.
One thing that irks me about banks is that checks don't appear on your balance sheet until they are cashed.

More and more people these days aren't balancing their checkbooks and simply using the bank website's readout of available money to determine whether or not they have enough money in their account to pay for something. However, this is flawed for checks as they aren't logged until cashed. I've had more than one occasion where I got burned by this thinking I had more money in my account than I actually did because I had sent out a check that my landlord didn't get around to cashing for a few weeks.

This seems like the way the world is headed and it still kind of boggles my mind as to why we have both the account ledger at the bank's website and a separate piece of software to balance our books. (Mint.com, Quicken, etc.) The only reason this dichotomy exists as far as I can tell is for things like checks which are outstanding and not debited against your account when you cut them. (Ok, obviously there are other things they do like budgeting, etc. but why can't that all be done with my bank's website?)

My ideal Bank 2.0 would incorporate things so if I pay my rent with a check or send a check to someone through the web billpay, it is deducted from my account immediately so I cannot accidentally spend that money which is out there. Obviously it would take user interaction to log the check they wrote to someone in the case of a personal check, but with the web billpay it could be done automatically. By reducing the balance when that check is cut, it should do it throughout the system (aka not simply symbolically) so if I go to the store and make a purchase with my debit card, my balance with the check cut is taken into account and it doesn't allow me to spend the money for the check.

In the event of a stop payment on the check or a lapse of a certain period of time, the money would be returned to my account, so a lost check wouldn't screw me over. This would eliminate this huge problem along with the pain of having to maintain an additional piece of software to monitor my account.

I've wanted the same thing, but I wanted to be able to view both numbers: the "real" balance and the "projected" one.
I think most banks already set aside the money from online bill paying services, though it may only be symbolically like you said. You have a good point about letting customers enter manual checks before they clear to help keep track of how much money you really have.
ING Direct debits the money from your account when they cut the check, and then re-deposits it (without interest) if it hasn't been cashed in 90 days or has payment stopped. They also don't issue manual checks, so that isn't an issue. Anecdotally, it seems to work well.

In regards to maintaining your own ledger instead of just using the bank's, the primary purpose is to help catch any errors that the bank makes with your account. Personally, I have few enough transactions that I can generally keep track of them in my head, but I could easily see how you might need some help with that if your finances were even a little more complicated than mine.

First I would cut the "Checks take N days to clear", "electronic transfers take 3 days", and other similar crap. Second I would allow the creation of multiple sub-accounts which can be used for single transactions, single receiver, or rate limited transactions. Think ING Orange sub-accounts but with external interfaces and on sterioids.
First I would cut the "Checks take N days to clear",

The only way to do this unilaterally is to loan your customers money you don't have for certain, yet. Legally, checks can be revoked up to six months later, so how many days a check takes to clear indicates nothing more than the confidence the bank has in final clearing, which won't happen for some time after that. So, people with bad credit, or who have previously overdrawn, are probably higher risks for depositing checks (especially personal checks) which may be revoked later.

(comment deleted)
I would eliminate overdraft fees on DEBIT accounts.
There is a law going into effect July 2010 which requires banks to ask you whether you want to be in the overdraft program.

If you decide not to be in it, your debit card will simply be denied at the point of sale.

If you elect to be in it, you will be allowed to overdraft your account, but will be charged a fee like now.

http://money.cnn.com/2009/11/12/news/economy/overdraft_fees/...

It does seem odd that almost anyone can get a no-fee credit card but it's hard to find a no-fee overdraft line of credit.
ING Direct and US Bank both have no-fee overdraft lines of credit.
I left Bank of America for that. It charged me $90 for over draft fees on my checking account for 3 cups of coffee, when I had a savings account linked to it with 5k in it.

Perplexing. I couldn't even stay on the phone long enough to hear their ridiculous explanation.

It's perhaps more a matter for regulation, but the current disparity between the value of the service and the cost should go. The customer should pay small, reasonable fees for everything. It is highly distortionary and is symptomatic of a dysfunctional market for banking services.

In the UK a few years ago the 'public debate' raged over ATM fees. Banks stopped cooperating and started charging about 2 pounds to the customers of other banks. It was pointed out that the amortised cost to them for a withdrawal was 30p. There was outcry and now they charge nothing. Apparently there are but two stable pricing policies - exorbitant and 0.

In general all basic banking is free, and they make profits by getting people into debt, or selling daft insurance schemes, or giving terrible rates of interest. Of course for the sensible individual this is not wholly a bad thing, if you can weather the hail of junk mail/promotions this business model entails and are prepared to move your savings account every year to make sure it has a reasonable return.

In my mind there is no reason for credit cards to exist at all - if you want a loan or overdraft, then get one - really they are a conflation of services: Visa/mastercard + overdraft + purchase insurance. What is desirable is a debit card that has visa/mastercard (encouragingly these do exist more and more, but are not the norm) and purchase insurance. i.e something functionally equivalent to a credit card, but without the capacity to get the owner into debt.

I like some of the other things people have suggested as well. Of course most of the ideas being voiced here are not exclusive!

Never call me on the phone. The promise of that alone would make me want to switch banks. Calls for security reasons are fine.

Seriously, the number of times my bank calls me is ridiculous. Usually it's to sell me extra stuff. Sometimes it's "to make sure my experience is satisfactory". It would be satisfactory if they'd switch to using email ;) Asking nicely to stop calling has not worked. Getting mad stops the calls for a few months. Then they start again. I guess their telemarketers are paid commission only.

Have an awesome web interface. Something like Buxfer, but built right in. In fact, I think Buxfer should become a bank :)

Marketing Idea - Make checks "Viral": Add "Deposit this check into a new account at X bank and receive an additional $100".

is for the disclaimer on back with legal requirements. $100 or $50 or whatever you think your new customer acquisition costs should be.

Different way of banking idea: Get rid of checking accounts all together. Instead you have a 2 Line of Credit accounts plus a Money Market account.

1 Zero interest LOC (ZLOC)and 1 variable interest LOC (VLOC).

At the end of the month the money in the MMA pays off the ZLOC completely assuming 100% coverage. Any overage goes to the VLOC.

Having the ZLOC eliminates the NSF fees a customer gets when they go over but it lets the bank generate interest fees on the VLOC if not payed by the end of the month. The deposits go to the VLOC first until it is zero, then move to the MMA.

Bank collects interest on MMA from the FED and makes money on loans fees/interest since you are bypassing the 10% reserve requirement and not even using the sweep method.

Marketing Idea - Make checks "Viral": Add "Deposit this check into a new account at X bank and receive an additional $100".

Just fwiw, this is a good example of how far behind US banks are. Until I moved to the US, I had barely used a check in the past 5 years. Now not only do I have to use them for a bunch of stuff (e.g. paying rent), I even have to pay for the checks themselves!

Banking 2.0 in the US should be less about adding features, and more about fixing all the blatantly broken stuff and dragging the whole system into the modern age, imho.

If ingdirect.com had paper checks (even for a fee) and a branch I could visit (if I really wanted to talk to a person) it would basically be perfect.
They're surprisingly good about answering the phone. No menu bullshit, just call and someone picks up.

If ING Direct interoperated with Mint, it would basically be perfect.

My bank 2.0 wouldn't be a bank as such, it would be an infrastructure provider with APIs for everything. Anyone could then build their bank-service on top of that. Want that debit card, this checking service and venmo.com for mobile micro payments? Just authorize them to your account, and you're going.

Any withdrawal from the account would require a full trace with timestamps and as much data as is available to be filed in the system. E.g. it would be expected for a web shop to save the invoice for your order directly in your account.

Still, I'd expect this to be regulation and liability hell, and if it isn't, I'm sure our friendly neighbourhood politicians and litigators will make it so in a jiffy.

Perhaps it's a lofty suggestion, but I would definitely bank at any bank partnered with a gold mint which allowed me to convert my savings to gold bullion with one-click. After the purchase, I would be allowed to visit my own safe with the physical bullion. That would be incredible.

Add a beautiful/easy online interface to manage your accounts, your gold, and foreign currencies.The bank would charge premiums for conversions in small quantities, or add a premium to the spot price.

Plus all the best aspects of current banks - as few or as low fees as possible, weekend hours, etc.

Check out pcfinancial.ca.

Here in Canada I thought we were totally behind on just about everything the states has. That was until I tried living there. Banking in the states is so fragmented and difficult! The big banks were a nightmare to deal with when something goes wrong and merely average when everything was fine. The smaller banks, some of which are awesome, are useless when you leave their prime territory. Even in Manhattan, it was difficult to find a branch let alone a bank machine of WaMu - which is the one I finally settled for. I'll admit that if you can get an employee type plan from Bank of America that cuts fees and makes it all nice and dandy then you are better off doing that.

OTHERWISE, look at pcfinancial. It's so basic and hasn't changed it's services much in the last 10 years I've used it. Use that as a base model and innovate. Also, look up how much money you need to have to legally register as a banking institution.

Give people interest on their savings.
Here, accountability is key: http://en.wikipedia.org/wiki/Accountability

So I'd use a social way to make all the structure accountable. A "real time" way of figuring out how any portion of the money deposited in the bank is being used, and the possibility to opt-out with one's money if, say, high executives have deep pockets or they invest in coal, oil, weapons, etc. (via funds, etc.)

No more overdraft fees. That's a start—a big start.