Tuesday, May 9, 2017

Fintech and Shadow Banks

"Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks" is an interesting new paper by  Greg Buchak, Gregor Matvos, Tomasz Piskorski, and Amit Seru

1. Shadow banks and fintech have grown a lot.
the market share of shadow banks in the mortgage market has nearly tripled from 14% to 38% from 2007-2015. In the Federal Housing Administration (FHA) mortgage market, which serves less creditworthy borrowers, the market share of shadow banks increased...from 20% to 75% of the market. In the mortgage market, “fintech” lenders, have increased their market share from about 5% to 15% in conforming mortgages and to 20% in FHA mortgages during the same period

2. Where are they expanding? They seem to be doing particularly well in serving lower income borrowers -- FHA loans.  They also can charge higher rates than conventional lenders, apparently a premium for convenience of not having to sit in the bank for hours and fill out forms,

Consider Quicken Loans, which has grown to the third largest mortgage lender in 2015. The Quicken “Rocket Mortgage” application is done mostly online, resulting in substantial labor and office space savings for Quicken Loans. The “Push Button. Get Mortgage” approach is also more convenient and faster for internet savvy consumers....
Among the borrowers most likely to value convenience, fintech lenders command an interest rate premium for their services. 
They also specialize in refinancing
Sector shadow banks have gained larger market shares in the refinancing market relative to financing house purchases directly. One possible reason for this segmentation is that traditional banks are also substantially more likely to hold loans on their own balance sheet than shadow banks. Approximately one fourth of traditional banks loans in HMDA are held on their own balance sheet. For shadow banks, the share is closer to 5%. Because refinancing loans held on the balance sheet cuts directly into a bank’s profit, their incentives to refinance are smaller..
This is a really cool point.

Our mortgage system is based on a rather crazy product, the fixed rate mortgage with a costly option to refinance. No other country does this. I know a lot of finance professors, and none of them can tell you the optimal refinancing rule. (It takes a statistical model of the term structure of interest rates and a complicated numerically solved dynamic program.) A lot of the system seems to be price discrimination by pointless complexity, a disease that permeates contemporary America.

Banks are on the other end of this. The bank holding your mortgage doesn't want you to refinance -- it wants you to keep paying the higher interest rate.  Unless, that is, it can get you to refinance too early and charge a lot of fees for it.  The natural product would be a automatically refinancing mortgage, in which a computer program automatically gives you a lower rate when it's time. It's not hard to figure out why banks don't offer that. In a competitive market, then, a third company would come in and offer refinancing, forcing the banks' hands. Competition is always the best consumer protection. And that seems to be exactly what we're seeing here.

3. Forces. A really good part of the paper (take notice economics PhD students) is how it teases out casual effects. I won't cover that in detail to keep the post from growing too long. A paper is not about its "findings" in the abstract, but the facts and logic in the paper. Some hints of the evidence follow.

To what extent are shadow banks and fintech stepping in to fill regulatory constraints, and to what extent is it just technology?

a) Some is technology, seen by this comparison.
Fintech lenders, for which the origination process takes place nearly entirely online... By comparing .. fintech and non-fintech shadow banks, we compare lenders who face similar regulatory regimes, thus isolating the role of technology. First, we find some evidence that fintech lenders appear to use different models (and possibly data) to set interest rates. Second, the ease of online origination appears to allow fintech lenders to charge higher rates, particularly among the lowest-risk, and presumably least price sensitive and most time sensitive borrowers.
b) The shadow banks primarily originate and then sell loans, and that business is practically all through government agencies these days. Private securitization fell off the cliff in 2008 and has not come back. 
In their current state, fintech lenders are tightly tethered to the ongoing operation of GSEs and the FHA as a source of capital. While fintech lenders may bring better services and pricing to the residential lending market, they appear to be intimately reliant on the political economy surrounding implicit and explicit government guarantees. How changes in political environment impacts the interaction between various lenders remains an area of future research.
In an otherwise cautious paper, I think this goes much too far. If a private securitization market existed, as it did before 2008, could shadow banks sell to them? Is the demise of private securitization just because the government killed it with the taxpayer subsidy implied by government guarantees? Absent guarantees would we just have a private industry that costs 20 basis points more? Just because finch now sells to government-guaranteed securitizers does not mean it must sell that way.

c) But the elephant in the room -- are shadow banks filling in where regulations keep transitional banks from going?
Unlike shadow banks, traditional banks are deposit taking institutions, and are thus subject to capital requirements, which do not bind shadow banks. If capital requirements are the constraint that increases the cost of extending mortgages for traditional banks, we should see larger entry of shadow banks in places in which capital requirement constraints are more binding. Indeed, we find a larger growth of shadow banks in counties in which capital constraints have tightened more in the last decade
In case you missed the point,
By comparing the lending patterns and growth of shadow bank lenders, we demonstrate shadow bank lenders expand among borrower segments and geographical areas in which regulatory burdens have made lending more difficult for traditional, deposit-taking banks.
"..the additional regulatory burden faced by banks opened a gap that was filled by shadow banks. "
We argue that shadow bank lenders possess regulatory advantages that have contributed to this growth. First, shadow bank lenders’ growth has been most dramatic among the high-risk, low-creditworthiness FHA borrower segment, as well as among low-income and high-minority areas, making loans that traditional banks may be unable hold on constrained and highly monitored balance sheets. Second, there has been significant geographical heterogeneity in bank capital ratios, regulator enforcement actions, and lawsuits arising from mortgage lending during the financial crisis, and we show that shadow banks are significantly more likely to enter in those markets where banks have faced the most regulatory constraints.
4. Policy

The paper is very careful not to make policy implications. I am under no such limitation.

It is too easy to take the last point and conclude "Regulations are hurting the banks! Get rid of them so banks can get their business back!" But that does not follow (which is a good reason the paper does not say it!)

Banks have capital and risk regulations because they fund their activities with deposits and short term debt. Those liabilities are prone to runs and financial crises. So in fact, one can come to quite the opposite conclusion:

The rise of fintech proves that there is no essential economic tie between loan origination and deposits or other short-term financing 

(Italicized because this is an important point at the end of a long post.) Maybe we want the crisis-prone traditional banking model to die out where it is not needed!

Update: Pedro Gete and Michael Rehr also find government-sponsored securitization helps the rise of fin-tech.


  1. “The rise of fintech proves that there is no essential economic tie between loan origination and deposits or other short-term financing”.

    I’m baffled. Any bank (regular, shadow or fintech) if it particularly wants to, can abstain from the “loans creates deposits” process. I.e. it can limit amounts it lends to whatever amount of money is supplied to it by shareholders, depositors etc. But there is no reason for banks to do that. I.e. if the money supply generally is expanding by X% a year, each bank might as join in the fun and create some money out of thin air and lend it out. That expands the amount of business each bank does.

    Assuming each bank does that, then there is a simple and unavoidable mathematical relationship between loan expansion and deposit expansion. That is, if private banks create and lend out $Y of new money, that money must end up as $Y of new deposits: it won’t vanish into thin air!

  2. I am a bit confused about this:
    "The rise of fintech proves that there is no essential economic tie between loan origination and deposits or other short-term financing"
    While many fintech companies rely on venture capital, my understanding is that some, like Quicken, also rely on wholesale funds and they are therefore subject to bank runs since they do raise funds by issuing short-term liabilities. Am I wrong?

    What I am getting out of this is quite the opposite: So long as the government tries to promote the goal of home ownership among low-income households, regulation aimed at limiting risk-taking by commercial banks will be ineffective since some other financial company will fill-in the void. Of course, we can say that what happens to them as a result of liquidity or interest-rate risk is their problem, except memories from the collapse of Bear Sterns and Lehman Brothers are still fresh. And of course, the default risk is still passed on to taxpayers who insure those loans, for example, through the FHA. Again, am I wrong?

  3. Fascinating post.

    I am reminded of the old S&L industry, protected for a while by Regulation Q, which allowed thrifts to pay 1/4 point higher on passbook accounts than commercial banks.

    Regulations that beget regulations.

    Yes, in the real world who would borrow short to lend long? Of course, there was the spontaneous creation of "building and loans" for a reason.

    Another oddity: In the old days, many thrifts were owned by their depositors! So it was really equity being lent out. Interesting option!

  4. Let's think this through in a mechanical way.

    Assume that money is a physical object. As a physical object, it must have an origination. Assume that banks (including central banks) originate money when loans are made (on a dollar per dollar basis).

    (Hence, money is destroyed when loans are repaid.)

    Now ask the question "Once created, can money be loaned?"

    Well, of course it can. In fact, how can it be loaned unless first created?

    Maybe when considering the creation of money, we are making a misleading assumption. Instead of assuming that all banks can create money simultaneous with loan creation, assume that SOME banks can create money simultaneous with loan creation. Then focus on which banks have that ability.

    Commercial banks generally have a restriction (enforced by government) that forces them to lend only from deposits on hand. That forces individual banks to have physical money in hand before a loan can be made. Thus, if a commercial bank makes a loan, the world can be confident that that loan has been backed by previously created physical money.

    Can we say the same thing about central banks? No, central banks are different creatures. Central banks can directly create money that has not been previously created. They can do this because government allows the practice. Acting through the central bank, government can borrow from itself. Of course, we say this in a kinder way; we say something to the effect that "A nation can borrow from itself."

    Hence, we rationalize the federal government printing money as 'just borrowing from ourselves'.

    How does this apply to shadow banking and this post? Mechanical logic would lead us to conclude that shadow banking depends upon the GSEs and FHA to underwrite the loans. Once the loans are underwritten by a GSE or FHA, they are adequately backed to enter the securitization market and be resold by the shadow bank.

    Closer investigation should reveal that all shadow bank loans find their way into the securitization market with GSE or FHA (meaning government) backing in some way.

    This is not a conventional way of rationalizing money and shadow banking, but where is it wrong?

  5. The definition of "shadow bank" is creeping. I am also not convinced that anything has really changed. For quite a long time the functions of mortgage origination, mortgage servicing and mortgage investing have become increasingly specialized activities.

    The mortgage originator underwrites the new loan and immediately sells it to the mortgage investor. If the loan is "conforming" (FNMA) or is FHA/VA (GNMA), the secondary market for mortgages is one of the most active and liquid in the world. Fannie Mae and Ginnie Mae exist to facilitate the transaction of loans between mortgage investors and enable the mortgage originator to be so specialized.

    What has changed? Countrywide went bankrupt. Countrywide was never a mortgage banker, they were a specialized (and the largest) mortgage originator. And while legally structured as a bank. Traditional banking -- accepting money from depositors using that to fund a pool of investments has never been their main line of business.

    The mortgage investor -- if they are a bank are tied to the rule that their investment pool is some multiple of deposits. The mortgage investor, however may not be a bank. It could be a pension fund, an insurance company, a mutual fund, etc. However, traditional banks are large investors in these mortgage pools.

    On the definition shadow banks. Prior to 2008, shadow banks were structured investment vehicles that bought the assets from traditional banks. Yes mortgages but also commercial loans, credit default swaps, and the least liquid assets on the banks books. These vehicles fit the roll of being investors not originators.


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.