Wednesday, June 24, 2020

Banks are slow to increase rates on savings accounts, but quick to reduce them

Chase Sunset & Vine, 2012. Painting by Alex Schaefer

There is a fundamental asymmetry to banking. Banks don't like to share higher interest rates with their customers who have checking and savings accounts. But they are quick to pass off lower interest rates to us.

This asymmetry is good for bank shareholders, but bad for customers.

To illustrate this asymmetry, I'll start by showing how banks modified interest rates on savings and checking accounts as the Federal Reserve, the U.S.'s central bank, went through a long period of hiking interest rates from 2015 to 2019.

The Federal Reserve's first rate increase (from 0.25% to 0.5%) was in December 2015. It increased rates once more in 2016 and three times in 2017. But the interest rate on the average U.S. savings account and interest checking account didn't start to rise till spring 2018, two and a half years after the Fed's first rate hike.

This irked me and I tweeted about it over a year ago:

If you're like me, you'd assume some sort of direct linkage between: 1) the interest rate that the Federal Reserve pays its customers (i.e. banks) and 2) the rate that these same banks pay their customers, you and me. Just like we have a checking account at a bank, banks maintain checking accounts at the Federal Reserve. They earn interest on balances held in those accounts. This rate is known as the Fed's interest rate on reserves, or IOR. As the Fed increases the interest rate that it pays on these checking accounts, the banks earn more from the Fed. But for some reason the banks are slow in passing these earnings on to the public.

Although the delay in pass-through irked me, I didn't take it too seriously, figuring it was due to some sort of institutional inertia. Banks are slow monolithic beasts. If they're slow to increase rates, at least they're slow to chop them, too, right? So on net, we customers aren't any worse off over the full economic cycle.
But if banks are slow to increase rates, is it indeed the case that they are also slow to reduce rates? Well, the results are in. The Fed began to cut rates in mid-2019, just around the time of my initial tweet. There were another few cuts in the latter half of 2019. Then COVID-19 hit in March, and the Fed rapidly ratcheted the rate it pay banks down from 1.6% to 0.1%. Banks went from earning around $38 billion in interest on their checking accounts at the Fed (in fiscal year 2018) to almost nothing.

If banks are generally lethargically about passing on rate changes to their customers, it should have taken them three or four years to reduce rates on savings and checking accounts back to where they had started. Nope. In just a month or two, the banks obliterated all the interest rate gains that customers with savings account had enjoyed since 2018:

So no, banks aren't lethargic beasts that are universally slow to change interest rates enjoyed by savers. They seem to have a strategy of increasing rates slowly, and then reducing them rapidly. Assholes.

Note that the savings rate I am using is from the Federal Deposit Insurance Corporation's website. FDIC takes the simple average of rates paid by all insured depository institutions and branches for which data are available.

By the way, this data probably doesn't represent the experience of the minority of financial sticklers who make an effort to locate high-interest rate savings accounts at online-only banks. JP Morgan's Goldman Sachs's Marcus currently offers 1.03%, much higher than the 0.10% that the Fed pays to Goldman JP Morgan. Ally offers 1.10%. But the average savings account holder doesn't bank at these institutions. They stick to Bank of America or Wells Fargo, which both offer a measly 0.01%.

This asymmetry is not a new phenomenon. In "Sticky Deposits", Federal Reserve economists John Driscoll & Ruth Judson found that rates are "downwards-flexible and upwards-sticky."

More specifically, the authors used proprietary data from 1997 to 2007 to show that interest rates on bank accounts and other retail deposits adjust about twice as frequently during periods of falling Fed interest rates as they do in rising ones. They estimate that this sluggish pass-through from rising Fed rates to customer rates costs American consumers around $100 billion per year!

My favorite chart from Driscoll & Judson is below:

Source: Judson & Driscoll

At left, we see the number of weeks it takes for banks to decrease the rate on interest checking accounts in response to a cut in the Fed's interest rate. At right we see the converse, how long it takes increase rates in response to higher Fed rates. Decreases tend to happen quickly (the purple bars in the left chart congregate closer to zero weeks) whereas increases are slow (the purple bars in the right chart congregate close to 100 weeks).

More specifically, during Fed easing cycles, checking deposit rates are updated on average every 22 weeks, but during tightening cycles it takes an average of 50 weeks.

So what explains this asymmetry? A lack of competition perhaps? If I had to guess, I'd say low financial education and dearth of customer attention. Banks can afford to be assholes because most customers either don't understand what is happening, or don't notice.

If the banks are taking advantage of their customers' ignorance and inattention to the tune of $100 billion per year, should something be done?

One option would be to provide a government savings option that 'corrects' for this asymmetry. Like digital savings bonds. Or maybe a government prepaid debit card with a built-in savings account. These cards would offer an interest rate that is linked to the Federal Reserve's interest rate, but only available to those below a certain income ceiling.

Or what about setting statutory minimum interest rates on savings accounts? In Brazil, for instance, banks are obligated to link the rate they pay on savings accounts to the central bank's interest rate:

Or maybe it starts with education. As part of its new financial literacy drive, Ontario will teach children how to identify Canadian coins and bills and compare their values in Grade 1, saving and spending from Grade 4, how to budget starting in Grade 5, and financial planning starting in Grade 6. If the result is a more savvy population, banks may face more pressure to pass on higher interest rates.

Or maybe nothing. In which case one hopes that over time the combination of better financial technology, branchless banking, and competition from Silicon Valley will eventually result in better pass-through and more symmetry in interest rates.


  1. Tx JPK, how does it compare with European banks ? p.d: Marcus is Goldman’s not JP

    1. Whoops, thanks.

      Here is a paper that finds the same asymmetry in interest rate passthrough for Europe:

  2. . I read that according to the results of the meeting of the Federal Committee for Open Markets, the Fed kept interest rates at the current level near zero, saying that it “doesn’t even think” about the increase. The central bank gave its cautious forecasts for GDP growth, employment and inflation. But the largest stock indexes at the close of trading on did not show a unified dynamics against the background of news about the Fed's interest rates at the same level.

  3. Does this article have a point? Banks are evil because they offer interest rates that are too low? And people are too dumb to know the options for seeking a higher return on their cash? Think you’ve been reading a bit too much Karl Marx. Love the solution is to offer savings rates linked to the Fed policy rate LoL. Pretty soon, the zero percent earnings these ppl have in their savings accounts will look pretty damn good relative to the “all knowing” Federal Reserve. What a joke.

  4. re: "an interest rate that is linked to the Federal Reserve's interest rate"

    LOL. The source of time deposits is other bank deposits, directly or indirectly via the currency route (never more than a short-term situation), or through the banks undivided profits accounts. Ergo, the banks collectively pay for what they already own. But that's not the worst of it. All 15 trillion dollars in bank held savings are un-used, un-spent, lost to both consumption and investment, indeed to any type of payment or expenditure. From the standpoint of the economy, the banks pay for their earning assets with new money - not existing deposits.

    The upshot is that money velocity falls, aggregate monetary purchasing power falls, AD, and thus N-gDp. That has been the case even before rates started falling due to the end of gate-keeping restrictions on savings accounts. That alone is responsible for Secular Stagnation, not robotics, not globalization, not demographics.

  5. Prima Facie Evidence. The 2018 pivot:
    As Dr. Philip George says: “When interest rates go up, flows into savings and time deposits increase.” (thereby destroying money velocity)
    The interest-bearing character of the DFI’s deposits which result in any sudden larger proportion of commercial bank deposits in the interest-bearing category destroys money velocity.
    2018-11-05 0.49
    2018-11-12 0.49
    2018-11-19 0.56 [spike]
    2018-11-26 0.57
    This is also an excellent device for the banking system to reduce its aggregate profits (as all savings originate within the confines of the payment's system, and an individual bank's primary deposit is a derivative deposit - from a system's perspective).
    It is hard for the average person to believe that banks do not loan out savings or existing deposits – demand or time. But the DFIs always create money by making loans to, or buying securities from, the non-bank public.
    This results in a double-bind for the Fed (FOMC schizophrenia: Do I stop because inflation is increasing? Or do I go because R-gDp is falling?). If it pursues a rather restrictive monetary policy, e.g., QT, interest rates tend to rise.
    This places a damper on the creation of new money but, paradoxically drives existing money (savings) out of circulation into frozen deposits (un-used and un-spent, lost to both consumption and investment). In a twinkling, the economy begins to suffer.
    % Deposits vs. large CDs on "Assets and Liabilities of Commercial Banks in the United States - H.8"
    Jul ,,,,, 12227 ,,,,, 1638.6 ,,,,, 7.46
    Aug ,,,,, 12236 ,,,,, 1629.4 ,,,,, 7.51
    Sep ,,,,, 12268 ,,,,, 1662.4 ,,,,, 7.38
    Oct ,,,,, 12318 ,,,,, 1685.8 ,,,,, 7.31 (twinkling)
    Nov ,,,,, 12313 ,,,,, 1680.1 ,,,,, 7.33
    Dec ,,,,, 12425 ,,,,, 1698.6 ,,,,, 7.31
    Jan ,,,,, 12465 ,,,,, 1732.9 ,,,,, 7.19
    Feb ,,,,, 12494 ,,,,, 1744.6 ,,,,, 7.16
    See: Dr. Philip George - October 9, 2018: “At the moment, one can safely say that the Fed's plan for three more rate hikes in 2019 will not materialise. The US economy will go into a tailspin much before that.”
    Or you could look at the Calafia Beach Pundit: “money demand fell from mid-2017 to mid-2018 as confidence soared and the economy strengthened”
    Link: September 25, 2018: “An Emerging And Important Secular Trend”

    Link: “Demand for money; what went up will soon come down”

  6. NSA N-gDp’s growth rates by decade, percent ∆:

    1970’s growth = 1.76
    1980’s growth = 1.15
    1990’s growth = 0.76
    2000’s growth = 0.52
    2010’s growth = 0.43

    Unless savings are activated, put back to work, a dampening economic impact, a deceleration in money velocity, is engendered and metastases, resulting in secular strangulation (not because of robotics, not because of demographics, not because of globalization).

    As the economic syllogism posits:

    #1) “Savings require prompt utilization if the circuit flow of funds is to be maintained and deflationary effects avoided”…
    #2) ”The growth of commercial bank-held time “savings” deposits shrinks aggregate demand and therefore produces adverse effects on gDp”…
    #3) ”The stoppage in the flow of funds, which is an inexorable part of time-deposit banking, would tend to have a longer-term debilitating effect on demands, particularly the demands for capital goods.” Circa 1959

    The answer to increased velocity, increased AD, increased incomes (the ingredient from which debt is paid), is to gradually drive the commercial banks out of the savings business. This will not reduce the size of the payment's system. It will make the banks more profitable.

  7. All monetary savings originate within the payment’s system. But a growth in time deposits depletes demand deposits by the same amount.

    Savers never transfer their savings out of the payments system in the first place, savings never leave the payment's system unless holders hoard currency or convert to other national currencies, e.g., FDI.

    In the context of their lending operations it is only possible to reduce bank assets, and deposits, by retiring bank-held loans, e.g., for the saver-holder to use his funds for the payment of a bank loan, interest on a bank loan for the payment of a bank service, or for the purchase from their banks of any type of commercial bank security obligation, e.g., banks stocks, debentures, etc.

    The NBFIs, e.g., hedge funds, insurance companies, pension funds and shadow banks are the DFI’s (regulated member banks), customers. The DFIs process all of the NBFI’s underlying payment transactions, both clearings, and settlements. The prosperity of the DFIs is dependent upon the prosperity of the NBFIs.

  8. Like Dr. William Barnett said (a former NSA Rocket Scientist).
    “the Fed should establish a “Bureau of Financial Statistics”. You can't duplicate this accounting today.

    1979: Double-entry Bookkeeping on a National Scale
    Loans and investments 1229.8
    Cash and Due from Banks 169.5
    Total Assets—Total Liabilities and Net Worth 1480.3
    Demand Deposits 400.5
    Time deposits 675.8
    Borrowings 180.5 (principally e-$s since 1969)
    Currency outside the banks 106.1
    Reserve Bank Credit 128.3

    MONETARY AND BANKING CHANGES End of 1939 to end of 1979 (figures in billions of dollars)

    (1) Net effect on the volume of time and demand deposits and borrowing of all factors, except commercial bank credit (principally capital accounts) 13.5
    (2) Net expansion of commercial bank credit 1189.1
    (3) Net increased in time and demand deposits and borrowings 1202.6

    Source: Computed from data reported in All-Bank Statistics, U.S. 1896-1955
    Federal Reserve; and the Federal Reserve Bulletin

    The fact is that from a systems' standpoint the banks pay for their earning assets with new money not existing deposits. This drastically changes everything in macro. For instance, the source of time deposits is demand deposits, i.e., the bank collectively pay for what they already own (very stupid and less profitable). So the domestic banks could undercut offshore lending, FX.

    M1 = currency outside the banks plus DD, including U.S. Treasury General Fund Account
    M2 = M1 plus all time deposits in the DFIs