Rising deficit in the US; rising surplus in China
The US released its March trade data yesterday. The US trade deficit widened. And as Menzie Chinn accurately notes, it wasn't just oil. The non-oil deficit also widened. Exports bounced back a bit from February, but the pace of growth still looks to me to be slowing. Y/y exports were up 9% or so. And non-oil goods imports jumped up — the y/y increase in March was close to 6%. Those numbers imply ongoing deterioration in the non-oil deficit.
The rise in oil imports stemmed more from an increase in the volume of imported oil than an increase the oil price — the March oil import bill was no higher than the January oil import bill, and remains well below its peak levels from last summer. The average price of imported oil was only $53 in March, just a bit higher than the average price in January (see Exhibit 17). The US oil import bill could rise further in April.
China also released its April trade data, along with data on its 2006 current account surplus. The April (customs basis) trade surplus came in at $16.9b — with 26.7% y/y export growth offsetting 21.2% import growth. The 2006 current account surplus was $250b — a rather large number, though one in line with my expectations.
The 2007 surplus looks to be substantially larger. If the April pace of increase for imports and exports is sustained for the entire year, the (customs) trade surplus would come in at around $270b, an increase of $90b from the nearly $180b 2006 surplus. Take the average y/y increase from the last three months (28.5% for exports, 16.2% for imports) and the surplus would come in at $325b, and increase of nearly $150b.
No wonder Stephen Green of Standard Chartered is now predicting that China's 2007 current account surplus will approach $400b, and its reserves could increase by as much as $550b.
The US data suggests that exports are still growing, just at a slower pace than in 2006. The fall in the dollar/ RMB over the course of 2006 seems to have done more to help China's exports than US exports.
Perhaps more importantly, it suggests that growth in non-oil imports has resumed, though here the trend isn't as obvious. Non-oil imports stalled in the second half of 2006, but the March number hints a renewed growth. See Exhibit 9.
All in all, the 2007 US trade deficit looks poised to at best stabilize and perhaps grow a bit if non-oil import growth resumes. That is disappointing, at least to me. It would be hard to think of a time when conditions globally are more favorable for reducing the US deficit.
US growth has slowed relative to global growth. That should help slow US demand for imports while increasing global demand for US exports.
The fall in relative growth rates isn't just a product of slower growth in the US either. Most other regions of the global economy are doing rather well right now.
- Europe has strengthened relative to the US, the euro (and pound) have strengthened relative to the US dollar and the US bilateral trade deficit with Europe is shrinking. If fell from $32.7b in q1 2006 to $23.8b in q1 2007, largely because US exports to Europe increased by nearly 21%. The US bilateral deficit with Canada is also falling. US imports from Canada in q1 2007 are 3.3% less than US imports from Canada in q1 2006.
- The oil and commodity exporters have dramatically increased their spending. Their collective surplus should fall dramatically this year so long as oil stays roughly where it is at – the same amount of export revenue and more imports should lead to a smaller surplus. Strong spending in Venezuela, for example, is one reason why the US bilateral deficit with Latin America fell from nearly $13b in q1 2006 to around $6b in q1 2007.
- India continues to do very well, spurred by strong domestic credit growth. The rupee has appreciated.
- China is obviously growing strongly. And while net exports are clearly adding considerably to the pace of China’s growth, the underlying momentum of growth inside China is strong. If the government of China took the breaks off banking lending, domestic demand growth would be even stronger. Right now, China’s government is restraining domestic demand growth to keep the economy – which is also getting a big stimulus from exports – from overheating.
- Strong growth in Asia though hasn't translated into a smaller US bilateral deficit with East Asia though. The bilateral deficit with the Pacific Rim was $86-87b in q1 2007 — up from $78b in q1 2006. US exports to the Pacific Rim are up 10.4% (q1 07 v q1 06), led by a 15.5% y.y increase to China. But US exports to fast growing China are increasing more slowly than US exports to old Europe. US imports from the Pacific Rim are up by about 10% — paced by a 19% increase in imports from China. Import growth from the rest of Asia was quite subdued.
The dollar has fallen – relative to all of Europe and a few countries in Asia (Korea, India), though obviously not all of Asia. And it would fall a lot more if the emerging world was intervening at an unprecedented pace …
The US has reduced its fiscal deficit. I am not a fan of the Bush Administration’s fiscal policy, but the Administration has not responded to the recent surge in corporate tax revenue with a new round of tax cuts or responded to the recent surge in income tax receipts from the top end of the income distribution with another round of tax cuts. Rather the recent surge in tax receipts has largely been used to bring the deficit down – it is now under 2% of GDP.
Basically, nearly all the conditions needed for adjustment are in place.
One is ingredient though is clearly missing. Many important Asian currencies remain weak. That presumably is why the US bilateral deficit with Asia is still increasing, while the US bilateral deficit with other regions is falling. It also helps explain the ongoing increase in both the Chinese and Japanese current account surpluses.
The yen is weak, even though Japan is now growing. But Japan isn’t (directly) intervening in them market: I think the shadow of past interventions helps support the carry trade, but that is another story.
And the yuan is weak. That is a direct result of government intervention. I am rather frustrated by stories that report the yuan’s cumulative appreciation against the dollar since mid-2005 without noting the dollar’s slide against other currencies since mid 2005. In real terms, the RMB hasn’t appreciated since mid 2005 – and in real terms, I think it has depreciated in 2007.
The RMB has moved by less against the dollar than most other currencies. The absence of more RMB appreciation is one reason why the US bilateral deficit with Asia isn’t falling. And China’s bilateral surplus with Europe is also rising fast – as one would expect, given how much the RMB has fallen v. the Euro. China’s overall surplus is clearly rising.
That is the problem, as I see it. Dollar weakness is leading the oil-importing portion of the dollar block to adjust. But it is doing more to push up China’s surplus than to push down the US deficit.
That means that the US is growing more dependent on Chinese financing. Like Stephen Green, I am looking for China’s reserves to increase by $500b this year, if not a bit more, and its purchases of US debt to top $350b. The formation of the state investment company might change the headline total a bit and reduce debt purchases a bit, but it won’t change the underlying dynamic.
I personally think the surge in Chinese financing of the US – and specifically the surge in indirect financing of US households (through Agency purchases, among other things) – helps to explain why the fall in the US fiscal deficit seems to have been offset by a rise in the overall deficit of US households, keeping the savings and investment balance from falling by as much as might be desired.
The lack of US household savings and the resulting US deficit seems to me to be partially induced by a set of policies that have pushed up China's surplus and as a result held down market interest rates in the US. I don’t buy Roach’s argument that the US savings deficit is independent of China’s savings surplus.
Clearly China isn’t alone in intervening to limit its currencies appreciation. But it is the biggest player, and the biggest constraint on more rapid appreciation by other emerging economies. I consequently share much of Fred Bergsten’s frustration with China’s currency policy.
China hasn’t done much to support global rebalancing over the past two years. Indeed, by following the dollar down, China seems to be actively retarding overall rebalancing. Rather than letting the RMB appreciate faster than the dollar depreciates, it has opted for more RMB weakness, more Chinese exports and more Chinese financing of the US. The dollar block is adjusting, just not in the way it should.
Part of the problem is that the existance of the dollar block, which links the currencies of the world’s biggest surplus country (now China) and the world’s biggest deficit country (the US) together too closely.
The data today probably with slow a broadly stable non-oil trade deficit in the US at a time when the US non-oil trade deficit should be falling. And it will probably show a further rise in the already large Chinese surplus.
Both results — if confirmed in the data — will be disappointing. Remember that a stable US trade deficit at current levels implies a rising US current account deficit over time. The interest bill on the United States external debt is — in my judgment – poised to rise substantially.
Brad,
By any measure versus the rest of the world, the United States is a high wage and cost of production nation. There are certain low value added, labor intensive industries that can never be globally competitive manufacturing in the United States; I don’t believe Americans are really very interested in making labor intensive shoes, or knitting sweaters and soccer balls for a living. Any US Industrial policy should emulate the German Economic model of high-value added, precision manufactured products without encumbering export restrictions. Former President Thomas Jefferson was explicit in stating that the United States should trade with all nations in the world, without entering into alliances or interfering in the internal affairs of other sovereign nations. If Thomas Jefferson were alive today, our President would be aghast at today’s Neo-con foreign policy which restricts US high-tech exports to over two-thirds of the world today, especially to potential fast growing markets in the Far East. A genuine return to free trade would open the door to greater American prosperity as the founding fathers would advocate.
China removes Capital Controls to permit citizens to invest $4.6 trillion in foreign stocks
http://www.bloomberg.com/apps/news?pid=20601080&sid=al8loG54lAfM&refer=asia
May 11 (Bloomberg) — China will let its banks buy shares overseas for the first time, diverting some of the country’s 35 trillion yuan ($4.6 trillion) of savings from a local stock market where trading has surged sevenfold.
Excellent blog.
Two minor disagreements.
“I think the shadow of past interventions helps support the carry trade”. This would be true only if JP CB signaled its unwillingness to see the Yen appreciate; but they havent done it, since many many years. And the levels of past intervention (102) are far from current levels. Why not open yourself to other explanations?
“I don’t buy Roach’s argument that the US savings deficit is independent of China’s savings surplus.” Of course: they are simultaneously determined. But how large is China’s impact? Since the US is almost the only country in the world with a large deficit, your constant reference to China to explain US deficits seems more like an obsession than a true explanation. But yes, Chinese policies are frustrating…
Unbelievable explosive rise today across the board in every sector for Chinese ADR stocks listed in the United States and Hong Kong. For instance, Guangshen Railway, an utility railway company from Hong Kong to Guangzhou, up 10 percent today. Explosive price rise today for state-owned China Mobile, Shanghai Petrochemical, PetroChina, Sinopec, China Life Insurance, Aluminium Corp of China, CNOOC, China Telecom, etc. In market capitalization, China Mobile is now the world’s largest telecom company, PetroChina is now world’s third largest energy company after Exxon and BP.
I think I will trash the sandwich and be going out for lunch today
one question i (still) have is what can the PBoC (or any other CBs w/ increasing dollar positions) trade it’s reserves for? by their own admission they’re beyond the point of having them for having-them’s sake and are seeking now to invest them in things other than debt, which just leads to the accumulation of more paper — IOUs that in DC’s estimation should be considered worthless.
DC is obviously frustrated by lack of access to US technology given current CFIUS rules (altho they’ve recently announced at least a $4bn commitment to purchase US tech goods). so if buying IBM, intel, apple & cisco (combined market cap ~$550bn) is off the table and they’ve reached their limit on dreamliners, what’s left that won’t give them, as BS sez, “very very low (probably negative) returns?”
well, so far they’re buying nuclear reactors from GE (and toshiba and areva) and other such “infrastructure” and, as DC is wont to point out, access to oil and mineral deposits around the world… but it also looks like they’re starting to expand their horizons and are buying real estate (a la 80s japan) and now apparently shares, if not majority ownership; in this kind of environment then, as jen has pointed out, the ‘asset class preferences’ of SWFs/SICs or whatever become enormously important…
…or boeing has reached its limit on delivering dreamliners, as it were
i think they are booked thru the next 4 yrs! (supposedly, due to ‘rational’ discipline, they’ve decided not to expand their production capacity…)
um, in other words, if i may chime in again, it looks like a reserve spending spree is on the way, this time for real assets instead of financial assets (with which they have had their fill); the question then i think is how much of it will bleed over into goods & services inflation and, of course, what the CB response will be to asset price inflation…
What can China spend its FX reserves on? A social safety net would be a superb place to start, so that households don’t feel compelled to save every other yuan that they earn….
While China has (finally) allowed equities to form part of the QDII program, bear in mind that a) overall quotas are still very small, and b) equities can form no more than 50% of total QDII assets, the other half being fixed income/money market funds that don’t sell. It is a step in the right direction, for sure, but a baby step, not a Neil Armstrong moon job….
so then, basically, does this just amount to hedge funds and PEGs having more competition? (or clients???)
From Reuters News Agencies, heavy buying of Chinese ADR stocks today by US fund managers to hedge against slowing US Economy
http://yahoo.reuters.com/news/articlehybrid.aspx?storyID=urn:newsml:reuters.com:20070511:MTFH76007_2007-05-11_16-21-27_N11439292&type=comktNews&rpc=44
NEW YORK, May 11 (Reuters) - Chinese stocks led advances in U.S.-traded shares of overseas companies on Friday as investors bet that the Chinese economy could be far from slowing down even as authorities work to keep it from overheating.
Chinese ADRs gained despite Beijing warning on Thursday it would keep tightening monetary policy to mop up excess liquidity and stop the world’s fourth-largest economy from boiling over.
Las Vegas Casino owners bet billions on China’s Macau
http://www.reuters.com/article/HotelsandCasinos07/idUSN1441895720070215
LOS ANGELES (Reuters) - The biggest bet in Macau is how fast luxury shopping and other non-gambling pleasures will take off in the Chinese casino enclave, and the Las Vegas players putting chips on the table agreed this week.
Macau gambling revenue has surpassed that of the Las Vegas Strip, and casino operators see more and more Chinese coming, but a key question is whether the shoppers heading for Hong Kong will make a detour.
Las Vegas Sands Inc. (LVS.N: Quote, Profile, Research plans to build more than 3 million square feet of shopping malls and luxury hotels, expecting the melange of activities that now drive Las Vegas will generate new business in Macau, rather than increased competition that would drive down profits.
Weidner likened Macau to 1960s and 1970s Las Vegas, but said casino operators this time around will want to cash in on non-gambling opportunities rather than cede the gains to other developers.
Macau has had gambling for more than a century but only recently allowed foreigners to invest, sparking a building rush. The country has been increasing the number of visas for mainlander visits, but travel is still restricted.
Gambling revenue rose 23 percent last year to $7 billion.
1. A rising deficit in America, yes, but nowhere near the (nominal) monthly highs hit in 2006. I will be gobsmacked if the 2007 US trade deficit exceeds the 2006 deficit, and I’ll bet most of you would be too.
2. Chinese stocks are into bubble-land with P/E ratios of 40-50x. I’ve been around long enough to see what happens afterwards. Prince and the Revolution sang about it. (Sorry, no YouTube clip available.)
To continue discussion from the previous thread since it remains relevant here:
I commented there that, if the US is not happy about China’s exchange rate policy, it seems to me that the most appropriate answer is to make it more difficult for them to buy the US financial assets they need to sustain the peg. Macro Man argues that this would be hard because they could buy through London etc, and keep them with a bank custodian etc. I wonder….these days, money laundering regulations make it difficult to anonymously hold even thousands of dollars, so it would be difficult for the Chinese to disguise ownership of their holdings. Just an idea!
recall PIMCO’s ZIRP remedy to global imbalances — http://www.rgemonitor.com/blog/setser/124806 — hard to sterilise when you’re getting zilch; ask japan! noone’s buying JGBs ‘cept the postal service
Interesting paper, guest-with-no-name!
Of course, PIMCO may have certain reasons for proposing such an idea!
On the contrary, paying zilch makes sterilisation easy!
RebelEconomist,
There is no secrecy or paradox to China’s PBoC purchase of US Treasury Bonds. The US government sells them, the Chinese government buys them. The US Treasury then records the Chinese ownership of the T Bonds.
DC, I am not arguing in favour of the US restricting Chinese dollar transactions; just saying that if they insist on doing so, it would be more appropriate to restrict the intervention, rather than the exports. While I do not fully understand the Chinese exchange rate policy (in their position, I would give more weight to the euro), as you know, I do not believe that the policy should be a problem for the Americans if they used the official capital inflows wisely.
DC, if you think that US treasury bonds are essentially worthless, why do you think the PBoC (or anyone else) continues to purchase them? because of ‘dollar hegemony’?
RE, am i missing something? you’re getting nothing and paying more (assuming domestic rates are positive) making sterilisation more expensive…
Emmanuel –
If Dr. Roubini is right and the US slips into an extended growth recession/ outright recession, tis hard to see the trade deficit getting bigger.
But if the consensus is right and the us economy resumes trend growth later in the year, that should pull up non-oil imports (there is already a bit of evidence of renewed growth).
Throw in a slowdown in exports (this is what really worries me) and oil above $60, and I could easily see the US monthly deficit well above $65b a month in the later part of the year. the full impact of the rise in oil prices in march won’t show up til april i suspect.
what will China buy, if not US debt — very very good question. hell if i know. I suspect that their foreign assets are growing so fast that they will buy about as much debt as before, but just throw in a few other things …
Wasn’t there a US TV gameshow some time ago before overconsumption became de rigeur where the “lucky” contestant was set loose in a large big-box retailer with an impossibly large budget, but a limited time frame in which to spend it? Teleport to the present, imagine trying to deploy $65billion-a-month in the real world! Even Buffett and Temasek are choking trying to deploy their paltry-by-comparison hordes. The leveraged real asset boom thus makes eminent sense in a world where the Chinese have sub-optimally chosen to first save their winnings in govt and mortgage collateral debt obligations, before perhaps spending them. Talk about telegraphing a trade!! For all leveraged real asset purchases, however impossibly low the earings yields or implied cap rates for the ones with yield, will, in this scenario, in hindsight, appear to be prescient front-running transactions by The Smartest-Guys in the Room who knew the Giant Stay-Puff Mashallow Man (remember from Ghostbusters?!?!) is turning the corner and about to embark upon mainstreet with the fattest wallet anyone has ever seeen.
To avoid this, the Smartest Guys in the Beijing Room will need first to engineer a whippingly big global credit-crisis and real-asset crash, such that they can pull the rug out from under the leveraged guys, and buy the real assets from will-be anxious distressed sellers, namely the leveraged bondholders, who will be regretting having lent to the Private Equity boys on such generous terms and conditions for such paltry spreads and will be looking for an exit amidst lots of others looking for the same….
Cassandra,
how would the Smart Guy in Beijing pull the rug? Stop buying US treasuries? Sell them? It seems this would be killing the goose that layed the golden egg; i.e. the dollar would plunge.
Guest-with-no-name,
I read the PIMCO paper and I was horrified at how bad it was….I actually checked the date to see if it was an April 1st spoof! If I had money with PIMCO I would take it back, because if these guys have any influence there then I would be worried about PIMCO’s investment competence, and if not, their fees must be too high!
Short of a repressive totalitarian state, you cannot just decree ZIRP. The central bank has to buy interest-bearing assets to lower their yield, and unless non-central-bank preferences make low interest rates feasible, the central bank would end up being almost the only lender, and hyperinflation would ensue. By helping US borrowers at the expense of lenders, the solution proposed would shift preferences in diametrically the opposite direction from what is required. I suggest that a bit of pain for US borrowers would be healthy…..I don’t mean contrived punishment, but just withdrawal of the protection paid for by the non-borrowers. Just about the only idea in the PIMCO paper I support is that the US needs more scientists. In particular, it needs more people who understand the principle of continuity from physics - ie the resources they use have to come from somewhere. If US consumers can become more realistic, they will prefer to borrow and consume less, and then US interest rates can come down in a sustainable way. They will buy less from China, who will then do less intervention to hold their peg, and the US current account deficit will come down (without a change in dollar/renminbi).
On sterilisation, what I mean is that in the case of China and Japan intervening against their currency, sterilisation involves selling domestic currency bonds. If domestic interest rates are low (because the preference for saving is high), then this sterlisation has little cost. In fact, unless the dollar depreciates faster than the interest rate differential, China and Japan will make money on their intervention. Of course, dollar deprecation is the $64bn question.
RE - On a technical point, ZIRP isn’t difficult for a central bank to achieve if it really wants to, ignoring the other potential consequences (currency, etc). There is always a degree of quantitative easing involved in any interest rate easing, nothwithstanding the formal distinction between the two in Japan. It’s easy for a CB to manufacture an excess reserve position for the banking system as a whole. Because excess reserves are an opportunity cost, banks can’t resist the idea of trying to spend their way out of their individual positions. They start at the short end of the curve, and as soon as they’re convinced the CB is serious, work their way out longer, buying bonds. The trick is that, although individual banks can attempt to work their work out of their own individual excess reserve positions, the system as a whole can’t do it so long as the CB runs an excess position for the system. The cat chases the tail continuosly so long as the CB wants it. Hence the continuing merry go round of banks buying securities and driving rates down. The end result is that the required monetization is done mostly by the banking system, not by the CB on its own
Of course, fears of hyperinflation might cause a mountainous kink at some point on the curve, if banks collectively feared this. (But this didn’t happen to a huge degree in Japan, for example). And it would be offset by borrowers shortening term on new issues.
Apart from that technical point, the paper appears to be seriously non-credible and outrageous from a policy perspective, and I agree with your general views re the paper and PIMCO’s credibility for publishing it. And you have a good point in that the hypothesis is so outrageous, banks might balk altogether at starting the chase - but not for long, once they acknowledge CB commitment and the carry profit they would forgo by not going along for the ride.
Guest-2,
Though my comments were in fact Friday-night cheek, they could perhaps begin by broadly buying credit default swaps across the market. THAT would seriously unsettle spreads, with knock-ons to equity risk premia, without slamming rates or the dollar. Flight to quality, rising deflation fears would bid up the prices of precisely the assets they own, and hit the prices of the real assets they perhaps want to buy.
agree ZIRP US is impractical, effectively monetising US debt, altho DC thinks it’s going to happen anyway (not believing the existing stock of US IOUs can be honoured), and it would make china think twice about purchasing US financial assets; tis an interesting thought experiment on the limits of US dollar hegemony…
jkh,
I admit that I find it difficult to understand how central banks can control interest rates at all (as do more acclaimed economists than me - see NBER working paper 7420), but it seems to me that to achieve ZIRP for any significant period without a massive expansion of the central bank balance sheet someone has to be willing to hold an awful lot of central bank money. If they are reluctant to do so, the result would presumably be hyperinflation. Japan could only sustain ZIRP because people were so cowed by the bust that they were willing to hold large amounts of central bank money. In America today, people are, if anything, overconfident about their finances. And if the central bank was able to achieve ZIRP in a confident economy, it is hard to see how it could be unwound without either massive losses to the public sector as the central bank sells its assets for less than it paid, or hyperinflation if the central bank refuses to sell its assets at a loss. You may be right that ZIRP is technically possible, but I hope that the US does not try it in present conditions!
I don’t condemn the PIMCO paper for its iconoclasm, but I do for its lack of detail and rigor in explaining how the proposal might work. Maybe some investors find such attention-seeking “research” impresssive though.
RE - thanks for referring NBER working paper 7420 “The Future of Monetary Policy” by Benjamin M. Friedman. I’ve read it, but disagree with its conclusions regarding CBs and domestic interest rate control.
Interestingly, one of the author’s arguments in the case of domestic interest rates is identical to one made recently on this blog in the case of central bank reserves and foreign exchange rates. One participant has argued (against considerable resistance) that central bank flows are immaterial in the bigger picture, and shouldn’t be attributed with undue influence in exchange rate determination. Friedman makes exactly the same point with respect to the effect of CB asset accumulation in the case of domestic interest rates. He maintains that central bank asset accumulation is too small compared to aggregate market flows, to have a significant effect on interest rates.
He also notes that financial innovation has resulted in the growth of non-bank money, non-bank credit, and private clearing systems (obviously true, but an old saw). He maintains CBs are gradually losing their monopoly position in the sense that more money is becoming non-reservable outside the banking system. He concludes: ” …the central bank cannot affect interest rates … in its country’s financial markets because borrowing and lending in those markets proceeds independently of whatever amount of reserves it chooses to supply …”
But neither the asset/flow argument, nor the fact that the ‘monopoly’ position of the central bank doesn’t incorporate the non-bank financial system, are relevant to the control that CBs exercise over domestic interest rates. And the case against the asset/flow argument if anything is even stronger for interest rates than for foreign exchange rates. It only matters that the CB retains its ‘monopoly’ with respect to the provision of reserves to the banking system. The banking system is part of the larger financial system. If asset prices are affected in the banking system, the effect will spread to the rest of the financial system through normal market transmission and arbitrage. If the CB wants to affect interest rates through reserve management for banks, the same effect will spread though the entire system. Banks will trade assets with non-bank counterparties if they become mispriced relative to the signals of CB reserve management. It is not necessary for a central bank to be involved in every transaction in the financial system in order to retain its ‘central’ influence.
The critical effect occurs via the level of excess reserves that the CB supplies to the banking system. This is a variable under the complete short-term control of the CB. (A more powerful tool than exists in the case of foreign exchange.) The CB supplies or withdraws excess reserves from the banking system in order to induce system behavior in buying or selling assets. There is an opportunity cost for individual banks when they hold excess reserves, and a penalty (financial and ‘moral suasion’) when they fail to meet minimum required reserves. Hence they will be induced to operate in such a way in the money markets as to square their positions against reserve requirements, thereby affecting domestic interest rates the way the CB intends. A CB using the leverage of system excess reserve management will induce the desired interest rate change through asset changes in the banking system, thereby lessening its own requirement for asset accumulation or reduction. Moreover, this leverage extends to asset pricing and asset trading in the larger (non-reservable) financial system through normal market price arbitrage and transmission.
The CB’s excess reserve setting has the effective leverage of a short-dated option, with a ‘strike’ equal to the underlying minimum reserve requirement. The excess setting is the true instrument of monetary policy, rather than the underlying reserve requirement. One practical demonstration of this is that the Bank of Canada eliminated required reserve levels some time ago, managing affairs quite adequately through the supply of reserves that are in effect excess to a 0 requirement, with penalties for individual banks being negative. Required reserves impose a form of tax on the banking system, but have nothing to do per se with the essence of monetary discipline - it is the excess or ‘option’ piece managed by the CB that does this.
Increased U.S. Export restrictions on Dual-use technologies to China
http://www.venturacountystar.com/news/2007/may/13/dual-use-technologies-vex-regulators-of-exports/
SAN JOSE The shaker machines built by Data Physics Corp. and exported to China can rattle locomotives. Or they can rattle missiles.
What they’re rattling is the central issue in the federal government’s case against the San Jose company and its China sales manager, one that highlights the difficulty in regulating the export of so-called dual-use technologies — devices with both military and civilian applications. Though subject to strict export regulations, their sale to China is legal if they are shipped to nonmilitary customers.
jkh,
Central banks do not have a (short run) monopoly of the supply of excess reserves. The alternative supply is banknotes.
Imagine a position in which the banking system balance sheet is in equilibrium at a certain level of interest rates, for a given amount of reserves. Let’s say that the central bank tries to raise market interest rates by restricting the supply of reserves (eg by raising the interest rate at which it will roll over a system repo at the prevailing interest rate). If the banking system passed this on, it could be expected to precipitate a contraction of loans and an expansion of deposits, reversing the textbook money multiplier circle and leaving an excess of banknotes. These then add to the banking system’s reserves as they are redeemed at the central bank.
Because the stock of loans and deposits is so large in relation to the amount of reserves, like Ben Friedman, I find it hard to believe that the return of banknotes does not overwhelm the central bank’s attempts to restrict reserves well before it is able to achieve its target interest rate.
And that is just the short run. If the central bank persistently attempts to use its monopoly to set interest rates at a level that does not clear the capital market, I would expect other forms of money to arise, the possibility of which motivated Friedman’s article.
While I find this subject fascinating, I appreciate that it is a bit off topic, so I would be happy to continue to discuss it with you bilaterally if no-one else is interested!
RE - I wasn’t clear on whether you wanted to pursue this discussion. (It started with ZIRP advisability/feasibility.) If not, just shut it down - your call. But given the lack of other traffic here now, here’s a (final?) response on one aspect.
You say, “Because the stock of loans and deposits is so large in relation to the amount of reserves, like Ben Friedman, I find it hard to believe that the return of banknotes does not overwhelm the central bank’s attempts to restrict reserves well before it is able to achieve its target interest rate.”
But it’s easy for a CB to offset the effect of banknote redemption. As you say, when a bank redeems notes with the central bank, its reserve settlement balance is credited. This increases system settlement balances with the central bank and results in an excess position in these balances, ceteris paribus. But the CB can easily offset this increase using open market or off-market operations. It simply sells assets to the market, or off-market to the government. The first transaction reduces private balances with the banks and corresponding bank settlement balances with the CB. The second transaction (for example, using swaps with the non-CB foreign exchange fund), reduces government balances with the banks (they do keep bank balances), and corresponding bank settlement balances with the CB. It’s a clean offset in either case, and a simple matter of consistent accounting across interlinked balance sheets (CB, banks, private, government).
Bank notes and bank settlement balances are both types of central bank liabilities. The whole point is that CBs closely control their asset-liability management, including the steps required to offset the bank settlement balance effect of bank note redemption. Accordingly, the amount of bank notes either in public hands or in bank inventories, and the issuance or redemption of these notes by the CB, is not a constraint on the CB’s ability to control the level of excess bank settlement balances.
The micro workings of this stuff are generally not found in economics textbooks, at least in my opinion. My own view of it, whether right or wrong in your view, is largely from personal experience in the area, which may partly explain why it is so different from the typical textbook formulation.
jkh,
I am glad to continue discussion, because it seems to me that this question of whether/how central banks control interest rates is a vital issue. If what central banks are doing is unsustainable in some way, it may be distorting financial markets, and possibly explain some apparent conundrums. No doubt Brad will let us know if he considers the discussion too far off topic (although it is has some relevance because reserve accumulation effectively amounts to using foreign exchange to supply central bank money).
You are right that the central bank can reduce reserves by another round of OMOs, but probably not at the rate they want to achieve. Whether they do a repo or an fx swap (aka a repo using fx as collateral), there is an interest rate involved. If the banking system can get the reserves they need with a smaller increase in interest rates than the central bank is trying to impose, then that is what they will do (except perhaps for some short bridging period before their slight increase in interest rates brings in the banknotes they need). Clearly the size of this problem for a central bank depends on the stock of banknotes in relation to need for reserves. I believe in most countries that do not pay interest on reserves, the stock of banknotes exceeds the need for reserves.
Of course, the central bank can sell some of the other assets backing its base money stock - eg treasuries from the SOMA in the US case - to reduce reserves, but the price they get is likely to reflect what they are trying to do, and may involve changing market prices that the central bank is not unhappy with (you may have noted my point at the end of the previous discussion that the amount of treasuries in the SOMA exceeds China’s holdings).
Ultimately, it seems to me that - without financial repression at least - central banks attempts to control interest rates is doomed to either failure or unsustainable distortion of financial markets. If I recall correctly, Michael Woodford has also considered this problem, and concluded that the market believes that the central bank has a good idea of what interest rate is appropriate, so that the rate the central bank sets works as a signal, rather than through market forces.
I wish I knew the answer to this question, but all I can do for now is to debate with people who say they do! Your participation is therefore much appreciated.
RE - I don’t know any topic with certainty; it’s more a strong belief in this case that the way it works is quite different than the typical textbook explanation. I appreciate your interest, and apologize in advance if my writing seems tutorial - it’s an attempt to flesh out one perspective. As you appreciate, it’s a very difficult topic from any point of view. I think we can assume Brad is OK with this conversation continuing at the tail end of this post unless we hear otherwise.
To be clear, I’m referring to the domestic reserves of a commercial bank and the related effect of a CB on domestic interest rates. The effect of CB foreign exchange reserves on domestic and/or foreign rates is an important but secondary aspect to the domestic interest rate issue. And the domestic rate is very relevant to the foreign issue, so I see it being relevant to this blog’s broad subject matter.
Also, I’m referring to the domestic rate spectrum over which the CB has the most control. My belief is that it has a very high degree of control, technically, over the overnight rate, and substantially less control and progressively less control, as one moves out the yield curve from the overnight rate. But it has virtually total control over the near-term path of the overnight rate - i.e., not necessarily every day, but as an average trend over the short term (e.g. fed funds rate). This is a technical view, assuming away other considerations that come into policy objectives, including the exchange rate. Prudent policy formulation is different than straight technical implementation of a hypothetical objective.
To avoid terminology confusion, I’ll use ‘DB’ for a domestic bank operating under domestic reserve requirements, within the sphere of a given central bank ‘CB’.
Different countries define required DB reserves in different ways, with different components and different tiers. Typically, the reserve asset mix will include some sub-range of DB settlement balances with the CB, bank notes held by the DB, and other eligible assets held by the DB.
First, the most critical component for domestic monetary management (in my view), and also the one that is specified and common to all CB/DB reserve relationships, is the set of DB bank settlement balances with the CB. This is a DB’s bank account with the CB, used for clearing purposes. It is through this account that the CB sets the real time system reserve status (e.g. excess reserves) and influences interest rates and monetary policy. This is the universal instrument of monetary policy by which a CB affects its banking system.
Second, CBs can affect the level of settlement balances by both open market operations (overt), and off-market operations (covert). The first method has both an intended reserve effect and an intended market signal effect according to the interest rate at which the CB trades with the market. The second method, very important but not widely understood, has an intended reserve effect without signaling any overt domestic interest rate intention. Domestic interest rates may be just fine, but the reserve setting may be off (for example due to bank note redemption), so the CB acts to recalibrate available reserves without perturbing domestic interest rates from their present position. Thus, the CB has full flexibility to adjust DB settlement balances with or without overt domestic interest rate signals.
Third, there is an enormous difference between the act of a CB market transaction at a given interest rate level, and the act of a CB in setting the level of available settlement system reserve balances. The market transactions are a direct signal about the CBs intended consequence for interest rate levels. The calibration of settlement balances is the bold stroke by which the CB prods DB behavior to move toward the desired interest rate target by natural market competition. A withdrawal of settlement balances prods competition for funds, driving rates higher, and an injection of settlement balances prods competition for assets, driving rates lower…(more on bank notes later)… OVER TO YOU
jkh,
We both seem to understand how monetary operations work (although I must say, despite having been a central banker myself, I had not heard much about covert operations!). I think that the issue comes down to one of scale. I am not an expert on the US monetary system, but it looks to me as if the banking system balance sheet is indeed huge (eg the stock of M2 seems to be about $7tn) compared to reserves ($20bn?) or even the entire balance sheet of the Federal Reserve System (
jkh,
We both seem to understand how monetary operations work (although I must say, despite having been a central banker myself, I had not heard much about covert operations!). I think that the issue comes down to one of scale. I am not an expert on the US monetary system, but it looks to me as if the banking system balance sheet is indeed huge (eg the stock of M2 seems to be about $7tn) compared to reserves ($20bn?) or even the entire balance sheet of the Federal Reserve System (less than $1bn). Loan demand and deposit supply don’t have to be that responsive to interest rates for the net effect of a modest change in the Federal Funds rate to totally overwhelm the capacity of the Federal Reserve system to adjust reserves enough to remain the marginal player.
You are right that the standard monetary economics textbooks do not cover the details of monetary policy operations. I do know of one good book on monetary policy implementation, by Ulrich Bindseil, but it does not provide an answer to my question though. It never ceases to amaze me how much opinion is expressed about interest rates (eg the PIMCO paper discussed in the previous thread) without worrying about the details of the mechanisms involved.
RE - ‘covert’ is my characterization. Here’s an example.
Most countries have their main foreign exchange reserve holdings separate from the central bank balance sheet. These foreign exchange funds are essentially assets of the government.
Suppose the CB wants to increase the level of DB settlement balances (a component of their reserves), but without resorting to open market operations. The CB can do a swap (essentially a reverse repo) with the FX fund. The CB cuts a cheque and buys an amount of FX from the FX fund, with an agreement to sell it back after 30 days, say. This increases the assets of the CB. The cheque is included as part of the government’s revenue receipts, which are deposited collectively in the government’s bank accounts with the DBs. Since this cheque was not sourced from a DB, its effect when credited is not only to increase the government’s deposit balances with the banks, but also to increase DB settlement balances on a total net basis. (In a normal DB bank clearing, settlement balances are exchanged but the net flow is 0, because every credit is sourced form a debit). Thus, system settlement balances are increased. These balances are a CB liability. The increase in assets of the CB (FX swap) has been offset by an increase in liabilities (DB settlement balances).
Now, because DB settlement balances have increased, they collectively have more money than they need to meet their statutory reserve requirement for the current period. Individual banks will attempt to rid themselves of excess settlement balances by spending their way out of their position. But the banking system as a whole can’t lower the level of balances, until the CB decides to withdraw what it has injected via the FX swap. The result of this is that DBs will collectively buy (i.e. monetize) money market assets and drive down yields.
The same injection/spending dynamic would apply in the case of settlement balances injected by open market operations. But the above is an example of an off-market transaction, or what I referred to as a ‘covert’ operation. The terms of the swap between two units of the government - the CB and the FX fund - may well be done at market rates, but are opaque to the market and therefore supply no transparent interest rate signal to the market at large. But the interest rate effect occurs indirectly and powerfully, via the swollen settlement balances of the DBs, giving an incentive to DB cash managers to buy money market assets and thereby influence market yields.
As I said before, this is not a well-known technique, but it is a very useful one.
Also, to repeat, the dynamics by which bank cash managers are induced to buy (or sell) money market assets, based on excess (or deficient) settlement balances, are very sensitive. It takes a relatively small change in system settlement balances for the CB to induce this behavior. The nominal size of the CB balance sheet or DB balance sheets or measures of money supply is really not an issue. The key is the leverage with which the settlement balance mechanism works - as I said it is analogous to option leverage - and the motivation that provides to bank cash managers to behave in a certain way. The CB uses this leverage to induce bank cash managers to blow up (or contract) their balance sheets with a corresponding effect on money market rates. The burden of affecting market rates by monetizing assets is transferred from the CB to the DBs. And the issue of scale is addressed by understanding the nature of excess reserve settings (i.e. excess settlement balances) as a highly leveraged monetary mechanism (like an option), to which bank cash managers respond very quickly.
jkh,
I think I understand your example, and I do not think it is going to overcome my objection, but before I set out my explanation, I would be grateful if you would answer a couple of clarifying questions:
(1) Do the government’s accounts with DBs bear interest?
(2) You said that you have personal experience of monetary policy operations….was this with the Bank of Canada?
RE:
1. It probably depends on the country, but in Canada the government balances earn interest. In fact, it’s a major issue of periodic negotiations between the banks and the government (interesting negotiations given that the big banks are essentially an oligopoly, and the government is the government - I’ve actually attended them). I don’t recall the exact formula, but it’s certainly something that is a market-oriented rate for such balances - there’s a lot of money at stake. I’ve never viewed this aspect as particularly relevant to the core issue of interest rate control by the Bank of Canada. But as you may gather, there are a lot of things that I view as peripheral to the crux of the issue, and only a couple at its kernel.
2. I managed the reserve position (including the Bank of Canada settlement account) and money market operations of one of the major Canadian banks, as one of my jobs there. As part of that job, I had ongoing communication with counterparts at the Bank of Canada, who were interested in how we were viewing the system, in terms of their monetary policy and our strategy to deal with it (including responding to system reserve settings and our interest rate views and money market strategies as a result). Central banks are always looking for information and feedback from market participants, and the big clearing banks are particularly important in this regard. This was some time ago, but the logic of the Bank of Canada’s balance sheet management hasn’t really changed, nor for that matter the essence of commercial bank reserve management. The environment has changed enormously in terms of the evolution from inflation to disinflation, the volatility of interest rates, and the integration of commercial banks with bank-owned investment dealers, but the fundamental way that the monetary system works really hasn’t changed. As I’ve mentioned before, the fact that the Bank of Canada eliminated reserve requirements some time ago also hasn’t changed the fact that it’s the excess DB reserve settlement balances (positive or negative, relative to a 0 requirement in this case) that are key to DB reserve management, money market operations, short term asset-liability management, and their effect on market interest rates (as intended by the central bank).
nice discussion — i wish i had enough time to really take in all in and join in at your level. the level of operational detail here is impressive.
jkh,
Given your experience, I am sure that you know how the central bank controls interest rates, and hopefully, if I ask the right questions, I will understand too. Actually, I think I have an idea how it works, but if I am right, the control is weak (and maybe wears off with time too), in which case US ZIRP now would not be feasible.
I should say that of course I accept that central banks, or indeed any other institution or even an individual with deep pockets, can control interest rates if they are prepared to lose money in the process. For example, if I am willing to lend at a lower rate than anyone else, I can certainly lower interest rates, and if I am willing to borrow at a higher rate than anyone else, I can certainly raise them, but of course, I am going to attract a lot of business, and unless I have an economic source or outlet respectively for this money, I will lose money on it. Call this the loss-leadership mechanism. If the government accounts did not bear interest, then that might have been an obviously important factor, hence my clarifying question yesterday.
I can see that the operation you describe adds reserves. It is rather like the exchange of government debt - in this case collateralised by fx - for central bank money that is the foundation of a fiat money system. I assume that settlement balances do not bear interest, in which case there must be a bit of either loss-leadership or financial repression going on here to get the DB to accept the government deposit. But having got the excess reserves, the DB wants to put them to interest-bearing work.
We then come back to the issue of scale. Let’s say that the central bank wants to lower the overnight interest by 25bps, and is allowed to reduce the interest on the government deposit by 25bps. I presume that the addition to the government’s deposit is small relative to the overall size of the DB’s balance sheet. Surely, if the DB cuts its loan and deposit rates by 25bps, loan demand increases, and deposit supply falls, by a combined amount far in excess of the increase in reserves. So, from your experience faced with this situation, why would the DB pass on the 25bp cut? Alternatively, why doesn’t a very large amount need to be deposited in the government’s account to make the 25bps stick?
RE-
It’s interesting to consider such thought provoking questions - it improves my own understanding of the subject as well. Its best to approach a detailed answer to your last question incrementally, so I’ll start from left field and 40,000 feet. This may take a few iterations at getting to your issue of scale more deeply.
If you may recall, we started out having a brief conversation on something like this a while ago. At that time, I mentioned the phrase “double-entry bookkeeping”. Another participant made the key point that there are many contingent (and perhaps dangerous) liabilities associated with a CB’s balance sheet in a fiat money system, and that you had to consider these in assessing the economic substance of a CB balance sheet. I agree with that view. But my objective then and now is to try and articulate how a central bank balance sheet works in real time, notwithstanding future contingencies of such a larger sort (such as hyperinflation or deflation for example). Expressed another way, I’m considering the short run implementation of domestic monetary policy as its assumed in our discussion - not the appropriateness of a particular CB policy in a fiat money system. Also, I’m considering hypothetically abrupt policy changes and associated hypothetical implementation changes (e.g. an abrupt change to ZIRP), which may not be advisable or even realistic, but in theory could be implemented. I’m not constraining this by dismissing it due to potential knock-on effects in disastrous policy ripples (e.g. currency crash, bond market collapse, hyperinflation, deflation, … etc.). I’m only addressing the question - can this be achieved technically by a central bank. Unconstrained by these other considerations, my answer is always yes.
I have several core beliefs on the subject. The first is that understanding the real time operation of a CB balance sheet requires a rigorous appreciation of related accounting entries for the CB balance sheet and DB balance sheets. This is necessary to understand the substance and meaning of the cash flows that are associated with monetary policy implementation on a daily basis. The second is that the domestic banking system is a ‘closed’ system from an asset-liability accounting perspective. More on this later, but what appears to be ‘open’ at the level of an individual DB, is closed at the level of the system. The CB constrains the operation of a closed system with its monetary tools and techniques. Third, the idea of leverage is essential to understanding the power that a CB wields over the DB system - the power, by managing relatively small changes in system settlement balances, to induce DB asset-liability management behavior consistent with a given interest rate policy and objective.
With that preamble, I’ll start into attempting to address your specific question in my comment that follows next.
RE -
Your most recent question was:
” Surely, if the DB cuts its loan and deposit rates by 25bps, loan demand increases, and deposit supply falls, by a combined amount far in excess of the increase in reserves. So, from your experience faced with this situation, why would the DB pass on the 25bp cut? Alternatively, why doesn’t a very large amount need to be deposited in the government’s account to make the 25bps stick? ”
Consider a simplified model where a DB balance sheet consists of CB settlement balances and loans (assets), deposits (liabilities), and equity. Expand this to the entire domestic banking system as {DB}, etc. The point is, if you hold settlement balances and equity constant, the net position of loans and deposits can’t change - every new loan creates a new deposit. It’s a closed system - loan proceeds are immediately credited to the customer’s deposit account. Changes in assets equal changes in liabilities. Given the model assumptions, as loans expand, so do deposits. I return to another comment I made earlier - most of the broad money supply (not the base) is created by DB lending. If loans increase, so do deposits. Insert into this a CB increase in settlement balances and government deposits. Again the change in assets equals the change in liabilities - settlement balances are {DB} asset; government deposits are a {DB} liability. Thus, the monetary base has increased by the increase in settlement balances, and the broad money supply has increase by the increase in government deposits. Both of these increases are small however, relative to the size of the {DB} balance sheet.
Implications:
a) Given the model, if the lowering of rates results in an increase in loans, it also results in an increase in deposits, for the entire system of {DB}. The macro mismatch to which you allude is not an issue. Any micro mismatch at the individual DB level is resolved through system competition for loans and deposits.
b) The DBs will attempt to preserve net interest margins through the lowering of both loan and deposit rates (although there can be peripheral technical issues such as interest rate risk mismatches). {DB} will tend to recoup on deposit costs what they incur in lost loan revenue. Government deposits will be a small part of the liability mix of DBs and incremental government deposits even less so. And interest rates will adjust down on these deposits according to formula. Thus, the interest rate cost of incremental government deposits is not very relevant to the dynamics of {DB} monetary response to a CB easing strategy - the issue is the effect on {DB} market activity from the injection of additional settlement balances (e.g. through the government deposit mechanism).
c) The issue of scale is addressed by the leveraged effect of settlement balances on {DB} behavior. I’ve mentioned this a few times, but probably haven’t explained it very well yet. I’ll work at it harder as we proceed. But to summarize for now, {DB} cash managers will compete to spend any excess balances provided by the CB every day that such excess balances are maintained. The {DB} system can’t spend the excess because it won’t disappear until the CB withdraws it from the system. Every marginal excess settlement balance that is spent on monetizing a new asset (say a treasury bill) will circulate back through the system to create a marginal new deposit every time. The balance sheet of the entire system of {DB} thus continues to expand every day, as the DBs buy up new assets every day, so long as they are stimulated to do so, by the system excess setting. If balance sheets expand, but margins are preserved, profits will actually increase.
RE:
I’m viewing ZIRP as a special, extreme case of the general topic of CB interest rate control. In the case of ZIRP, customer loan rates will be at spread to ZIR, and customer deposit rates will be at a lesser spread, so a positive net interest margin will still result. There will still be a yield curve on Treasuries, so DBs buying assets in the market will benefit from the curve. This isn’t to say there isn’t pressure overall on margins as the general level of overnight rates heads to ZIR. Such happened in Japan and put pressure on Japanese bank margins. But I would think of ZIRP as an extreme limit of what we’re generally discussing here, which is interest rate control by the CB. It does present challenges, but it can be done.
RE -
Here’s a ‘Bank of Canada 101′ blurb that I picked up from the web. It’s probably oversimplified, but consistent with what I’ve written.
Cash Management
The major method for influencing interest rates is the Bank’s management of cash deposits in the banking system. Every day, the banks try to balance all cash flowing in or out of their accounts as closely as possible to avoid a positive balance in their clearing account at the Bank of Canada that earns no interest or a negative balance that incurs interest charges. The Bank of Canada can affect these balances by moving cash between federal government accounts at the Bank and demand accounts at the commercial banks. An increased positive balance tends to lower interest rates while an increased negative balance tends to raise rates.
Open Market Operations
If cash management does not achieve the desired effect, the Bank buys or sells 3-month treasury bills in the open market. A purchase, for example, is paid for by crediting the account of the seller at the Bank. This process is in essence printing money. The money supply is increased and the commercial banks have additional funds to expand their credit lines.
Moral Suasion
On rare occasions, the Bank has advised the chartered banks to alter a particular lending activity, to increase or decrease their lending activity or to work towards the desired level of interest rates.
RE -
Here’s a more sophisticated paper from the Bank of Canada website:
http://www.bankofcanada.ca/en/review/winter05-06/chant.pdf
Towards a Made-in-Canada Monetary Policy: Closing the Circle
John Chant* 2006 Bank of Canada review
Some excerpts:
” … the Bank relied on transfers of government deposits to manage bank liquidity, thus developing developing a monetary policy instrument that bypassed the money market …while a significant portion of these transfers represented the neutralization of the liquidity effects of government receipts or disbursements, there was also a monetary policy component whereby movement of deposits added to bank liquidity by increasing bank claims on the Bank of Canada, while transfers from the banks to the Bank had the opposite effect. The Bank’s use of this technique was distinctive in that it turned management of government-deposit balances into anactive instrument to bring about changes in bank liquidity…
… long recognized as a tax on banking services, the cash reserve requirement was phased out between 1992 and 1994. Instead of holding positive reserves, banks were expected to maintain zero clearing balances over the reserve-averaging period. At the same time, incentives to meet the zero cash requirement were strengthened by balancing the costs of holding holding deficits and excess balances…
… In 1994, the Bank added clarity and emphasized its focus on short-term rates by adopting an explicit 50-basis-point operating band for the overnight rate, the limits of which were reinforced …though this range was not publicly announced (as a fixed Bank Rate would be),changes in the range would quickly become apparent to market participants through observing the Bank’s operations in money markets. The Bank made the target range for the overnight rate still clearer in 1996 when it returned to fixing the Bank Rate, setting it as the upper limit of the operating band for the overnight rate …
… the Bank also revamped its approach to reserve management: the level of clearing balances was to be maintained at roughly zero,typically confining government deposit transfers to neutralizing the impact of public sector flows. Arrangements for government deposit transfers for preceding-day value were replaced by same-day settlement. As well, given the fact that Canadian banks knew with certainty their positions at the end of each business day and had a period to trade surpluses and deficits with each other before final settlement of their LVTS (large value transfer system) clearing balances at the Bank of Canada, the need for a reserve-averaging period to smooth fluctuations was eliminated …”
jkh,
Right; what you describe is the textbook money multiplier cascade of lending and redepositing.
The next issue to resolve is that the money multiplier depends on, in addition to the {DB}’s reserve ratio, the non-bank private sector’s currency to deposit ratio (not all textbooks explicitly identify this). It seems likely that this depends on interest rates. If so, then the money multiplier falls, potentially (unless it falls very little) cancelling out the effect of the increase in reserves. In other words, since the opportunity cost of holding currency falls, the non-bank private sector draws out more banknotes, and, with very little fall in interest rates, the increase in the {DB} settlement balances is drawn on to pay for banknotes.
Is this not what happens, in your experience?
RE -
I’m not describing the textbook multiplier. It’s something very different. That’s the whole point. What I’m describing is not based on required reserves. In Canada, it can’t be, because there are no required reserves, which is why it’s a good illustration of what I’m trying to describe. What I’m describing is a different dynamic. It’s a process by which DBs can lend excess settlement balances ad infinitum, in theory, until the CB withdraws them from the system. This has nothing to do with a required reserve ‘multiplier’.
For the Bank of Canada system, the textbook multiplier obviously fails, because there are no required reserves. For systems with required reserves, it also fails. Here’s one reason why. In required reserve systems, there is a time lag between the data point for reservable deposits and the specification of required reserves, and the implementation of that required reserve setting. The deposit data collection must precede the corresponding reserve requirement period. This has to be the case - the clearing banks are too big and the system is too big for it to be manageable otherwise. There’s no way that a data collection system for the entire banking system can provide real time feedback on deposit changes as they occur, with corresponding reserve requirements calculated such that reserve requirements can be managed on a co-temporal basis with deposit changes. The deposit data collection is completed, and reserves requirements are then set for the subsequent reserve period. Because of time lag, the CB will typically provide required system reserves when requirements change. This is a passive response to reserve requirements as determined by deposit balance changes.
Thus, reserve supply follows deposit creation, not vice versa. This may sound radical to you, because it is the opposite direction of causality from what you are taught in textbooks. But it’s the way it works. What I’ve been describing, again, is excess reserves. In the case of the Bank of Canada, it’s the excess level relative to a required level of 0. Whatever the system, if the CB doesn’t like broad money supply growth and the fact that deposits and corresponding required reserves are increasing, it will squeeze excess reserve levels to drive up interest rates, in an attempt to dampen loan and deposit growth. So whether there are required reserves or not, monetary conditions are steered by the CB setting of excess reserve levels.
The excess reserve deployment I’ve described for the Bank of Canada, with no reserve requirements, is the same as would be the case for a system with required reserves. That’s how the Bank of Canada could get away with eliminating reserve requirements. And that’s why required reserves don’t matter for monetary management, and that’s how CBs manage monetary conditions through excess reserve settings. Required reserves are a tax. They are not required for monetary management. The textbooks have it wrong. The ‘money multiplier’ is in truth only a statement of a relationship between deposits and reserves. It’s got nothing to do with the true causality of DB/CB monetary management.
This is a critical point obviously. I guess I’m trying to convince you of it, but it’s looking less likely that you want to go there. I’ll leave the bank note aspect to later. It’s a minor aspect to the main theme and not important when one believes in the practical irrelevance of the textbook multiplier. At the same time, there’s little fruit in discussing it while there’s an impasse on the main point. Perhaps I’m not explaining it clearly. The problem is that it’s hugely different than the textbook explanation. Just ask yourself, how can the Bank of Canada, with no required reserves, manage things according to the textbook explanation?
jkh,
I know what you mean…….I grew up in a system with no reserve requirement too! But you don’t need required reserves to have a deposit multiplier. The textbook idea is that a bank keeps some central bank money as reserves for liquidity purposes. The multiplier reflects prudential bank balance sheet management practice. All you need is some penalty or cost of not having enough reserves. Presumably, you did not aim to lend out all your bank’s central bank money - you kept back some, either vault cash or a reserve in the settlement accounts, no? And if you did, that amount was related to the size of your bank’s balance sheet, particularly its sight deposits?
RE -
Unfortunately, I disagree with everything in your last.
The classic textbook multiplier has to do with required statutory reserves, not voluntary prudential reserves. Point me to anything that states authoritatively otherwise.
Central bank money is not for prudential liquidity management. It is for bank settlement clearing.
And yes, excess settlement balances are lent out. What remains is negligible in a 0 reserve requirement system, and what is mandated as minimum in a non-0 reserve requirement system.
Prudential liquidity management is over and above statutorily specified reserves, involves maintaing interest bearing liquid assets, and certainly does not involve maintaining excess settlement balances.
jkh,
A typical British textbook will not assume reserve requirements in explaining the money multiplier, because we do not have them. My copy of Begg, Fischer and Dornbusch, for example, does not. Even Mishkin’s US-focussed Economics of Money, Banking and Financial Markets considers the possibility that reserve requirements may not be binding.
Note that I included currency (in bank vaults) as well as settlement balances in central bank money. I know that settlement balances were always very small in the UK before they were renumerated, but a bank’s currency holdings at least would be increased if the bank grew its balance sheet, wouldn’t it?
But maybe the extra currency required by a bank as its balance sheet grows is so small that it does not represent a significant impediment to balance sheet expansion. If not, and loans made return as a deposit elsewhere in {DB}, what was it that limited the expansion of your bank’s balance sheet?
Maybe the textbooks are so unrealistic that they get in the way of understanding how interest rates are controlled. What was the mechanism that made your bank pass on the change in interest rate set by the central bank?
Anyway, I will sign off for today. Feel free to give up trying to resolve this if it is taking too much time. Thanks for your patience so far!
RE -
Thanks for the multiplier info.
I agree with you that DB inventories of notes are central bank money in addition to settlement balances, and that currency holdings would increase as bank size grows. And the amount would be fairly modest relative to the rest of the balance sheet expansion.
The primary force limiting the expansion of a bank balance sheet in my view is the allocation of bank capital to various businesses in the bank, also an interesting subject. Liquidity management and capital management are quite strategic. Management of CB settlement balances and money markets is a more tactical exercise.
” What was the mechanism that made your bank pass on the change in interest rate set by the central bank? ” I’d say that it was our version of the old maxim, ‘ Don’t fight the Fed’. What a CB wants a CB gets, including its desired overnight interest rate. If there’s market resistance, or doubting of the telegraphed rate change, the CB will simply lean into it by exaggerating the adjustment to {DB} cash balances, prodding market activity toward the intended rate change. The intended interest rate effect is transmitted from CB settlement balances and related {DB} activity to the rest of system via pricing arbitrage, if discrepancies exist. Discrepancies don’t last for long in a liquid market. Anybody that stands in the way of the intended rate move will get creamed. It just happens that way, when the CB decides that it will happen.
I don’t want to give up trying to resolve these issues, but a break might be best at this stage. I’ll be out shortly anyway, for 10 days or so, so it’s a good time to pause. It’s been an excellent discussion. I appreciate your willingness to listen, and your feedback and ideas. Let’s pick it up again some time.
See you in a future post!
jkh,
Sure, thanks for the discussion. I have been trying to work this question out for several years, so I am sure that I will remain interested if we return to it later.
I said I think I have some idea how it works. I suspect that a combination of factors allow central banks limited control of the particular interest rate that they focus on. Among them are:
(1) the demand for currency is highly inelastic with interest rates, so that the multiplier is fairly rigid
(2) in practice, being guided by macroeconomic models, central banks do not try to set an interest rate that is totally out of line with what Wicksell called the “natural rate of interest” and Adam Smith called the “rate of mercantile profit”
(3) central banks can afford to provide some loss-leadership, since their seigniorage profit is relatively painless to give up
(4) they are able to repress private sector banks to some degree - ie your “they can make us do it if we refuse” point above
My key conclusion is that, if central banks push it too far, like trying to implement ZIRP in the US now, they would generate serious market distortions and would probably have to give up.
Regards
like quantum physics, sounds fascinating, but way over my head
esp wrt ’signaling’ how perception helps determine reality…
i’d still like to know tho hypothetically (but in practice) — as in they never would, but if they did — whether the US could implement a ZIRP!
and fwiw, they’ve been thinking about Monetary Policy Alternatives at the Zero Bound when all the world’s like switzerland
cheers!
RE, guest -
This is still bothering me. I’m going to try and construct a simple numerical model, comparing the textbook multiplier to my interpretation of the same problem,and post it by the end of the week. I’ve actually never tried to do that. It will be a test to see if conceptual intuition can follow through with numerical logic. (Maybe I won’t be able to do it, and will surrender). RE, I haven’t quite buckled yet to the force of your knowledge and argument.
jkh,
The aim is not to make you buckle! I want to find the answer, and since you did the job, you probably know it, even if sub-consciously.
RE -
As per my previous, examples follow of a neutral CB reserve setting, and a tightening CB reserve setting, and how the dynamic differs from the textbook multiplier.
I’m off for 1 week. I could construct examples of CB easing, including CB ZIRP, after my return, if you are interested. The aspect of central bank notes also remains to be done.
Regards,
Example 1 - Neutral CB
Suppose a given domestic banking system includes a CB and commercial banks, {DB}, collectively with $ 1 trillion in deposits. Assume the required reserve ratio is 5 per cent of deposits. Then total required reserves for the system are $ 50 billion. Assume the system holds $ 10 billion in bank notes, which count toward reserves. The remainder of reserves is held in the form of settlement balances with the CB, which for the entire system therefore must be $ 40 billion.
A ‘reserve period’ is defined as the time period for which a particular reserve requirement is in place. The requirement is calculated according to the deposit count in the preceding period. In theory, the period could be as short as 1 day, with new reserve requirements coming into force each day, based on the deposit count for the previous day. In practice, the periodicity likely varies among different country systems. But there must be some discrete time lag between the time of the deposit calculation and the enforcement of the corresponding reserve requirement. There can’t be instantaneous continuity in such a function in practice.
So assume that in period 1, the CB provides the system with required settlement balances and no excess reserves. Suppose the CB is generally comfortable with monetary conditions. Suppose {DB} extend new credit of $ 2 billion, thereby creating deposits in the same amount. This results in future reserve requirements of $ 100 million, to be in force for the next period. Total required and actual settlement balances of $ 40 billion remain unchanged for the current period. The CB takes no action yet because it is comfortable with this credit and money expansion.
In period 2, the new reserve requirement has increased by $ 100 million. Still content with monetary conditions, the CB supplies these newly required reserves to the system in the form of increased settlement balances. Total system settlement balances are $ 40.1 billion. The CB remains comfortable with the pace of credit expansion and overall monetary conditions. It supplies no excess reserves and undertakes no open market action to change the overnight rate.
Thus, the CB supplies the required reserves resulting from deposit expansion, so long as it is comfortable with monetary conditions.
The actual causality is the inverse of that of the textbook multiplier. The CB supplies new reserves in response to new deposits - {DB} does not generate new deposits as a result of new reserves. The actual dynamic is not a {DB} reserve multiplier, but more a CB deposit ‘fractionater’. The CB reserve response to expanding credit and money is passive so long as it is comfortable with monetary conditions.
Example 2 - Tightening CB
Assume the same starting point - {DB} with $ 1 trillion in deposits, $ 50 billion in required reserves, $ 10 billion in bank notes, and required system settlement balances of $ 40 billion. Also, assume the same expansion of credit and deposits in period 1. The new required settlement balances for period 2 again are $ 40.1 billion.
This time, the CB has become uncomfortable with monetary developments in period 1 - e.g. bad inflation data. It decides to tighten monetary conditions in period 2. Instead of supplying required reserves, it withholds $ 500 million from the new required level, resulting in an actual level of settlement balances of $ 39.6 billion instead of the required level of $ 40.1 billion.
Notwithstanding the collective shortfall in available reserves, individual banks still seek to meet their own requirements. This results in a scramble for funds. They begin to call loans and compete for available deposits. This puts upward pressure on market interest rates. Liquidation of credit results in a decline in system deposits (changes in bank assets are offset by changes in liabilities, ex changes in equity). This puts further pressure on deposit competition and rates. The process continues until the end of period 2. Despite their individual efforts, banks collectively still have insufficient reserve levels at the end of the period. Some individual banks are forced to borrow from the CB to meet their requirements. There is a financial penalty for doing so.
Suppose at the end of period 2, the CB is comfortable with higher market rates that have resulted from the {DB} competitive response to the tightening program. Suppose as a result of this response, credit and deposits each declined by $ 4 billion in period 2. This reduces the reserve requirement for period 3. Assuming no change in notes, the required system settlement balances are now $ 39.9 billion. Since the CB is comfortable with market rates, it supplies the required reserves for period 3, increasing available settlement balances from $ 39.6 to $ 39.9 billion. Thus, the CB successively supplies, withholds, and then supplies required reserves in response to its view of overall monetary conditions.
The period 2 dynamic is quite different from the textbook multiplier, which prescribes that bank(s) respond to a reduction in reserves by shrinking deposits in order to balance with the reserve ratio. That’s not what is happening here at all. Bank(s) are not attempting to reduce their deposit levels. On the contrary, individual banks are attempting to increase deposits in an attempt to restore required reserves, given the system shortfall in supplied reserves. The actual contraction in system deposits is the unintended consequence of banks seeking to restore reserves by calling loans. The CB withdrew $ 500 million in required reserves, which was its judgment of the amount necessary to elicit the intended interest rate effect. The drop of $ 4 billion in system deposits was 8 times the reserve decrease, unrelated to the conventional multiplier prescription of 20 times (1/.05). The reserve decline was temporary, only lasting as long as necessary for the CB to achieve its desired interest rate result. Accordingly, the deposit decline was completely arbitrary, based on the duration of the tightening, the degree of competition for funds, the resultant credit tightening, and the associated path of interest rates. The deposit response was not a {DB} strategy to reduce deposits proportionate to the reserve decline. The deposit response was entirely market driven, not reserve multiplier driven. In period 3, the CB continues its passive reserve response to both the deposit base established in period 2, and the subsequent path of credit and money, so long as it is comfortable with monetary conditions.