Monday, February 1, 2016

Bank of Japan warms up the potato



Will the Bank of Japan's negative rates work?

Many people say no, among them Louis-Phillippe Rochon:
Sadly, they won't. They [negative rates] are based on a faulty understanding of our banking system. The reason banks do not lend is not because they are constrained by liquidity, but because they are unwilling to lend in such uncertain times.
Banks lend in the hope of getting reimbursed with interest. But banks are too pessimistic about the ability of the private sector to honour their debt, and so prefer not to lend. Having extra cash courtesy of the central bank imposing negative rates won't change the dark economic narrative. [link]
I disagree. Even if the lending channel is closed, a negative rate policy still sets off a hot potato effect that gets the Bank of Japan a bit closer to hitting its inflation targets and stimulating nominal GDP than without that same policy.

For the sake of argument I'll grant Rochon the point that negative rates might not encourage banks to lend. And as you'll read in the comments here, that would certainly have implications on the effectiveness of monetary policy. But even if we close the door on loans, the interest rate cut will simply find a different route into prices and the real economy.

The moment the BoJ reduces the rate on deposits it creates a hot potato; an asset with a below-market return that its owner is desperate to be rid of. Bank reserve managers will simultaneously try to sell off their BoJ deposits in order to get a better return in short term corporate and government debt. In aggregate, however, banks cannot get rid of reserves, which pushes the prices of these competing short-term assets up and their expected returns back in line with the return on balances held at the central bank, a process that continues until reserve managers are indifferent on the margin between owning BoJ deposits and short term corporate/government debt.

The hot potato doesn't stop here but continues to cascade through financial markets. At the margin, corporate and government debt will now be overvalued relative to other financial assets (like stocks), encouraging fund managers and other investors to bid up the prices of all remaining assets in the financial market until returns are once again in balance.

Up till now the the hot potato that I've been describing has been trapped in Japanese financial markets thanks to Rochon's blocked lending channel. Acting as a bridge into the real economy are the portfolios held by consumers. Japanese consumers own not only portfolios of financial assets but portfolios of consumption goods that yield an ongoing flow of consumption services. Think cars, shavers, tables, and vacations (the latter of which yield a recurring flow of memories). Likewise, financial assets yield an ongoing flow of consumption services since interest payments and the final return of principle can be measured in terms of consumption. When prices in financial markets rise and returns fall, a portfolio of financial assets now yields a smaller discounted quantity of future consumption services than a competing portfolio of consumption goods. In response, consumers will re-balance out of financial assets into undervalued consumption goods, causing consumer prices to rise. Or, if there is some stickiness in prices, the quantity sold experiences a boom.

And that's how the hot potato ignited by the Bank of Japan's negative rates gets passed into consumer prices and the real economy when the lending channel is closed.

---

Another interesting critique of the effectiveness of negative rates has to do with the fact that in those nations that have already experimented with negative rates, the penalty has not been passed through to retail deposits. This could be a problem because if Japanese retail depositors are not going to be fined by banks, that nullifies the hot potato effect I described above. After all, consumers won't bother trying to re-balance out of the financial economy into the real economy if they can just hoard superior-yielding 0% deposits.

This failure to pass-through negative central bank rates will probably not be more than a short-term phenomenon. As the BoJ deposit rate get ever more negative, those banks that choose to prop up the rate sthey offer on retail deposits allow themselves to be the victims of arbitrage, consumers taking the positive end of the deal as they migrate into superior-yielding deposits. Borrowing at 0% to invest at -0.1% isn't a particularly profitable place for a bank to put itself in. The only way for a bank to rectify the situation is by the passing-through of negative rates to retail clients or the setting of limits on retail account sizes. The hot potato effect gets new life as investors flee bank deposits by purchasing underpriced consumer goods.

As Gavyn Davies points out, the Bank of Japan has set a tiered negative rate whereby the full effect of negative interest rates is not felt by banks; only a portion of deposits held at the BoJ will be docked the full 0.1% while the rest get off Scot-free. This BoJ (i.e. taxpayer) subsidy to banks helps offset any financial losses that banks incur by choosing to avoid passing through negative rates to retail customers, thus encouraging bank managers to keep retail deposit rates steady at 0%. .

But as the BoJ continues to cut rates, the size of the BoJ subsidy is unlikely to increase as fast as the size of the penalty imposed on banks as measured by the gap between the cost of maintaining 0% retail deposit rates and the revenues earned on negative-yielding central bank deposits & other short term money market assets. To plug these growing losses, and absent a compensating subsidy, banks will have no choice but to pass-through negative rates to retail clients or put a limit on retail account sizes. This in turn will give free rein to the hot potato effect.

Negative interest rates are like water, they'll always find a crack.

35 comments:

  1. JP, great to see this one.

    "Will the Bank of Japan's negative rates work?"

    Is was my question to you, Sumner and Rowe.

    I don't know if this is still true, but does this have any influence on your thoughts:


    "Japan, say financial experts, is still a “cash society.” Visitors from the United States or Europe are surprised at how little use the Japanese make of credit cards and how much cash they carry around in their wallets. Yet Japan is an economically and technologically advanced country and, according to some measures, ahead of the United States in the use of telecommunications and information technology. So why do the citizens of this economic powerhouse still do business the way Americans and Europeans did a generation ago? The answer highlights the factors affecting the demand for money.

    One reason the Japanese use cash so much is that their institutions never made the switch to heavy reliance on plastic. For complex reasons, Japan’s retail sector is still dominated by small mom-and-pop stores, which are reluctant to invest in credit card technology. Japan’s banks have also been slow about pushing transaction technology; visitors are often surprised to find that ATMs close early in the evening rather than staying open all night."

    Krugman and Wells Macroeconomics p452

    ReplyDelete
    Replies
    1. In other words is there much room for increase in cash held by the public, if that factors in to a channel?

      Delete
    2. I think that means that the Bank of Japan may have less room for deeply negative rates than, say, the Riksbank. Japanese seem to be more comfortable with cash than Swedes, which means they may be quicker to bolt out of deposits into paper as the deposit rate is reduced.

      Delete
    3. But perhaps there's a simple relation between the public's holding of paper and the price level (more so than the existence of base money (MB) in general). Check out these four plots.

      All four have had their parameters adjusted for a best fit to the data. The plot in the lower right (the best fit to the data of the four) uses MB - reserves (i.e. cash, or "M0" for short) as the measure of money. It also uses a slowly varying parameter "kappa" (rather than fixed, like in the plot in the upper right, in which a constant kappa is fit to the data). For the slowly varying kappa (lower right), kappa = log(M0/c)/log(NGDP/c), where c is a constant fit to the data. There's actually a hypothesized reason for the various functional forms, which I won't get into.

      But the point is a "bolt out of deposits into paper" may actually be helpful if increased inflation is what's desired (assuming the relations in that plot holds). Unfortunately, the Japanese public may have already done that to a large degree, so there may not be much room to increase that (and thus not much room for inflation to increase).

      Here's a plot for Japan corresponding to the one in the lower right above (i.e. the same hypothesized relation between CPI and M0 and same hypothesized expression for a slowly varying kappa). Only in this case the data and the model (for "P" or CPI) are the two gray curves on the bottom, while the blue curve is M0.

      Delete
    4. There aren't many parameters to fit: one (in the case of the the QTM (which does the worst of the four)) and two for the others.

      Delete
    5. What I find intriguing is that there's no model of human behavior there: no utility maximization, agents, micro-foundations or expectations. It just assumes human economic behavior is random (or complex enough to appear random and intractable); an idea explored by economist Gary Becker back in 1962.

      Delete
    6. An increase in the amount of currency doesn't cause higher price level; a higher price level leads people to draw out more currency.

      Delete
    7. OK, but is it possible that the a higher price level requires there be more currency available to draw out?

      In other words, could lack of currency to draw out prevent the price level from rising?

      Delete
    8. Well the Bank of Japan doesn't control the quantity of paper currency in circulation. It passively adapts to the demand for that currency. If people want to hold their wealth in cash rather than deposits, it will print whatever quantity of paper is necessary to meet redemption requests.

      Delete
  2. Unless.. it persuades everyone to expect deflation and that even with negative rates it will still hold its value. If applied to large accounts but not small ones, it could even be beneficially redistributive.

    ReplyDelete
    Replies
    1. The yen did move in the right direction at least.

      Delete
  3. If BOJ buys everything what is the price of the last asset, higher or lower than market return? If less what is the amount of purchases that turns the table so that marginal effect is negative? If higher why socialism hasn't worked anywhere?

    ReplyDelete
    Replies
    1. Sounds like a Nick Rowe post. Not sure what this has to do with negative rates though.

      Delete
    2. I think lower interest rate can be translated to mean more asset purchases as the BOJ needs more liquidity to support lower interest rates?

      Regardless whether you dismiss my translation I would be interested in your opinion on asset purchases in general. It seems that asset values are boosted by QE but then again I doubt that the last assets on the list would be worth much given the lesser incentives (socialism). Something I have often thought and think you would have an informed opinion on.

      Delete
    3. I tend to think that once the lower bound is hit, asset purchases only have limited effectiveness as a monetary policy. After that, a central bank needs to reduce its deposit rate or do forward guidance if it wants to be effective.

      Delete
  4. For consumers its possible low returns on financial assets may not increase current purchases but decrease current purchases where the consumer aims to maintain the future purchasing power of their savings by spending less.

    ReplyDelete
    Replies
    1. Well given my setup--expensive consumption in the financial markets vs cheap consumption in the consumption goods market--investors are choosing to ignore a profit opportunity.

      Delete
    2. JP, agree about the consequences of your setup, but disagree that your setup is useful. The evidence from China is that financial repression (artificially low returns on savings) decreases consumption. Call it an intertemporal budget constraint or not but it's there, and it becomes more binding the more worried consumers are about the future.

      Delete
    3. So let's say its 1995 and the Fed reduces rates from 5% to 4.5%. You don't think that would be stimulative? (My setup applies just as well to positive rate environments where banks are impaired as it does in negative rate ones).

      Delete
  5. I have some serious doubts about your hot potato effect. It would seem to me that return expectations are often positively correlated with recent price action.

    If I have a painting on my wall and a similar one sells suggesting the valuation of mine is $10m higher I am most likely inclined to leave it right on the wall. I see no mechanism in which that affects aggregate demand unless there is some physical item I need and don't already have. Unlikely if I have modern art on my walls already.

    Just an example, i know. But i think the link between financial asset values and aggregate demand is much looser and less straightforward than you would like to believe.

    ReplyDelete
    Replies
    1. Take Microsoft shares. Microsoft has a set of underlying earnings that provides an earnings yield. When its share price get marked up, all else staying the same its earnings yield declines so that fundamental investors expect a lower return. (Technical traders might have different expectations given the positive correlation to recent price action, but at the end of the day fundamentals dominate). In hunt of larger returns, fundamental investors may look to the consumer goods market--and that's how you get the link between financial asset values and aggregate demand.

      Delete
  6. Seems like this should be a cold potato effect. Start with your financial market effects. All short-term rates of interest go down as a result of the reduction in the interest rate on reserves. Now currency - a zero-interest asset - looks more attractive. It's a cold potato that people want to hold on to longer and not spend. The price level goes down, and in the dynamic process the inflation rate must fall.

    ReplyDelete
    Replies
    1. I agree that the existence of cash means we get some sort of cold potato effect, but only when rates go negative enough. Denmark and Switzerland have dropped rates to -.75% but thanks to storage costs we haven't seen an increase in cash holdings.

      Delete
    2. Couldn't a Bank exchange its diminishing reserves for a non diminishing Deposit account by opening a deposit account with another bank and making a large deposit. Theoretical but possible.

      Also are we going to see the end of the use of the term "the lower zero bound"

      Delete
    3. Why would another bank offer a non-diminishing account to a competitor? It would be unprofitable.

      I think the term ZLB is being replaced by ELB, where E is effective.

      Delete
    4. ELB is not descriptive either - the rate at which currency becomes more attractive is different for different economic agents. There's not a single lower bound threshold. It's a distribution, and it's not a fixed distribution.

      Delete
    5. When you say that it's a distribution, aren't you just describing any old demand curve, except this is for cash storage services?

      Delete
  7. I am of the opinion that most of the money in bank deposit accounts belongs to business and to pension funds. I don't know how to support that opinion.

    I know that business and local government needs large amounts of cash to meet reoccurring expenses. This cash will reside on deposit in banks.

    I know that pension funds and stock brokers hold enormous amounts of cash deposited by clients. This must reside on deposit in banks.

    Retail customers are unlikely to hold much cash in banks. First, a large number of retail customers exist near poverty level. Without research, I would throw a number of 30% of the U.S. population into that category.

    The remaining 70% would not hold much cash on deposit at banks. They might place cash into the hands of stock brokers but we have already identified stock brokers as likely holders of cash.

    I suppose that pension funds and stock brokers could increase their fees to recover the tax paid due to vastly increased negative interest rates. Would that induce holders of pension fund and stock broker accounts to spend their money instead of saving it? I suppose it would at some high level of negative interest rates.

    There must be some tension between borrowing to own a house and investing in the stock market. May people do both which seems to me to be equivalent to borrowing to enable pension and stock purchases. (Borrow at home to allow funding a pension and stock speculation.) That activity would be discouraged with very high negative interest rates.

    Thanks for the well considered post.

    ReplyDelete
    Replies
    1. Roger, I agree with your intuition that most money in bank deposit accounts belong to businesses rather than retail.

      i.e. see Miles Kimball here: https://twitter.com/mileskimball/status/692563675685752832

      Delete
  8. Why wouldn't Japanese savers withdraw their savings from banks and flood into gold in order to avoid a negative retail deposit rate? At a low enough rate (-2%? -3%?), it will still be worth the storage costs and risks of gold to go into gold. There's a store of value that will not just help you dodge the negative deposit rate penalties; it will also gain in yen value (and dollar value, for that matter) as the yen (and dollar) are inevitably forced to depreciate for political reasons as Central Banks try to avoid an inevitable recession. Over the last 10 years, gold has gone from $600/oz. to $1200/oz. It has gone from 60,000 yen/oz. to 140,000 yen/oz. Why on earth would any Japanese saver buy a 10-year JPY government bond at a 0.05% annual rate? (Much less hold onto yen at -0.1%?) It is insane! Money capitalists are going to revolt and stampede into gold if they try to push this too far, and the Japanese Central Bank will be left as the only one buying Japanese government bonds with printed money, at which point you can expect hyperinflation.

    ReplyDelete
    Replies
    1. "Why wouldn't Japanese savers withdraw their savings from banks and flood into gold in order to avoid a negative retail deposit rate?"

      Japanese savers would indeed withdraw into gold as it allows them to dodge the negative rate. But the price of gold will quickly spike until it reaches a level at which it is expected to decline at the same rate as the negative interest rate. That is what equilibrates the market. After that, the gold exit is effectively closed and depositors will look to other exits like the consumption goods market.

      Delete
  9. Quick thoughts:

    “Another interesting critique of the effectiveness of negative rates has to do with the fact that in those nations that have already experimented with negative rates, the penalty has not been passed through to retail deposits.”

    Do you include fee effects?

    How do you know this?

    “But as the BoJ continues to cut rates, the size of the BoJ subsidy is unlikely to increase as fast as the size of the penalty imposed on banks as measured by the gap between the cost of maintaining 0% retail deposit rates and the revenues earned on negative-yielding central bank deposits & other short term money market assets. To plug these growing losses, and absent a compensating subsidy, banks will have no choice but to pass-through negative rates to retail clients or put a limit on retail account sizes. This in turn will give free rein to the hot potato effect.”

    I’m not sure how correct this is.

    The immediate interest margin concern for the banks is the change in the rate they earn on reserves. And it looks to me like the bulk of the reserves will pay an interest rate of zero, thereby being immunized against the negative policy rate. It is only a relatively small amount of marginal excess reserves that will earn a negative rate. That will certainly spill over into wholesale money market rates more broadly. But the money market effect should be contained and absorbable on the balance sheets of the banks. The real question is how the change in the risk free rate will affect return on capital hurdle rates and pricing for loans. That may be stickier (risk premia partially expanding to allow for a declining risk free rate), allowing deposit rates more broadly to be stickier. And then there are fees.

    ReplyDelete
    Replies
    1. I'm sort of contradicting myself in those two points, but just wondering in the first how firm the data is.

      Delete
    2. "How do you know this?"

      It's just what I've been reading in the financial press. Which could be wrong.

      "I’m not sure how correct this is."

      Well, let's say that the majority of the reserves earn 0% so that as the BoJ drops from -0.1% to -0.3% to -0.6%, banks don't really lose much from that source since only the marginal reserve deposit is being taxed. The central bank's negative deposit rate also determines the nominal return on a bank's short term assets, too. And that's where the bank's spread really starts to narrow if it chooses to keep retail deposit rates at 0%, no?

      Delete
  10. "The central bank's negative deposit rate also determines the nominal return on a bank's short term assets, too. And that's where the bank's spread really starts to narrow if it chooses to keep retail deposit rates at 0%, no?"

    Yes - in terms of the short term risk free rate that banks use to anchor the pricing of their floating rate retail products

    so more right than not in that respect

    but I'm thinking of "net stickiness" in retail asset pricing for one thing

    e.g. the kind of sluggishness demonstrated by Canadian banks in the last bank rate drop - where I believe they only lowered their prime rate (and floating rate mortgages based on prime) by 15 basis points (or was it 10) in response to a 25 basis point bank rate drop

    also, fees on the liability side

    banks would freak out if they started publically posting negative rates for retail deposits - because their customers would freak out a fortiori

    what I'm saying also is that the pricing transmission for wholesale money market rates may start to diverge from that for retail rates

    and there's also the issue that the transmission effect to retail assets and liabilities covers a larger balance sheet scope than the scope for transmission to money market assets (apart from the reserve account) - which is what I meant by absorption capacity - eating a negative spread on a relatively small portfolio compared to the asset-liability scope of the associated retail pricing decision

    but there's a limit to just how much the banks could withstand depending on the determination of the central bank - at some point with a negative enough policy rate they'd have to cave for sure

    which I guess is also your point

    ReplyDelete