As if previous studies were not clear enough (see also here for a primer by Justin Merrill), new evidence has just been added to the rapidly growing body of research that demonstrates the very limited effectiveness of macro-prudential measures. MC Klein, on the FT Alphaville blog, points to a new research paper by Shin, Bruno and Shim, and to a 2011 report by the central bank of Spain, which both clearly highlights the flaws in central bankers’ and regulators’ current economic micromanagement worldview.

As Klein says, the Spanish assessment of macro-prudential measures is “not encouraging”. Spain is one of the rare European countries to have actively used macro-prudential tools (dynamic provisioning) in the period preceding the crisis, without having any control on monetary policy. The effectiveness remained limited as we have noticed.

But it gets worse. Shin’s new paper covers 12 Asian countries that have also extensively used macro-prudential policies over the past decade or longer. Results are more than mixed, to say the least. What do they find? (my emphasis)

Our findings therefore highlight a fundamental question in the rationale for macroprudential policy. To the extent that monetary policy works by intertemporal allocation of spending, loose monetary policy encourages greater borrowing to bring spending forward from the future to the present. Macroprudential policies work by restraining borrowing. Our empirical findings suggest that the macroprudential policies have been employed so that they pull in the same direction as monetary policy – that is, macroprudential policies are introduced during periods of monetary tightening. First, the correlations are especially high between monetary policy (interest rate policy) and banking sector CFM and domestic macroprudential policies. The correlation is lower between monetary policy and bond market CFM policies, but this finding likely reflects structural shifts in the capital markets of the countries in our sample.

Second, when we measure the monetary policy stance with the Taylor rule gap (ie the difference between the actual policy rate and the Taylor rule rate), we find that non-interest rate monetary policy tools were used in a complementary way with monetary policy during the first phase of global liquidity (pre-2007) after controlling for global liquidity and country-level variables. Last, our study also suggests that when monetary policy and banking inflow measures are pulling in the same direction (opposite directions), banking inflow measures are successful (not successful) in slowing down foreign investment in domestic bonds. Such a conclusion is also consistent with the principle that when monetary policy and macroprudential policies pull in the opposite direction, economic agents are being told simultaneously to borrow more and borrow less.

This is clear. Unlike what central bankers would conveniently love to believe, monetary policy remains key. Liquidity always finds a way. And macro-pru ‘leaks’ as Aiyar, Calomiris and Wieladek said already some time ago. Basel’s risk-weights make it more capital efficient (i.e. cheaper) to use extra liquidity for real estate lending purposes? Let’s introduce targeted macro-pru tools to prevent too much liquidity from flowing into this sector. Fine. But now, perhaps corporate lending, or equity investing have thus become more attractive from a profitability perspective. Market participants constantly adjust their P&L calculations. Don’t expect them to simply sit on an ocean of unused liquidity simply because you said so. They’re going to use it. Perhaps not on their first or second investment choices. But if the risk-adjusted yield they can extract generate a positive net present value (implied: vs. the opportunity cost of sitting on cash), they will invest. Worst, a long period of artificial surplus liquidity has the potential to suppress risk-assessment: investors panic and attempt to extract any yield from any sort of assets. Macro-pru just transfers the problem. Monetary policy, as Jeremy Stein from the Fed once noted, “gets in all the cracks”.

The authors also warn against a phenomenon I had remarked in a previous post on macro-prudential tools, and which is symptomatic with studies based on regression:

To the extent that new macroprudential policies happen only after a period of discussion within the government, central bank and other public authorities (such as financial regulators), the introduction of such policies often coincides with the late stages of the boom. To the extent that the boom subsides under its own weight, the introduction of the macroprudential policy and the subsequent slowdown of capital flows and credit growth would be a coincidence, not a causal effect. To this extent, the results reported below should be taken with some caution.

In the end, their results question the very validity of current regulatory thinking: “don’t worry about interest rates, we have a macro-prudential toolkit to deal with excesses and financial imbalances”. Well, apparently not. Their reasoning is that business and financial cycles are separate and not synchronised (see Vitor Constancio, from the ECB, here, or virtually any central banker speech over the past 3 years). In fact, those cycles look highly synchronised and very likely to simply be different sides of the same coin, as economists such as Mises, Hayek or Minsky have theorised. Implying that changes in asset and credit markets have no economic consequences and can be dealt with separately is dubious.

An increasing number of studies are now available. Some, like the one in this post, even originates from formerly advocates and designers of macro-prudential policies. The ball is now in the policymakers’ court. (although some don’t seem to care much: remember that BoE’s Mark Carney declared that the BIS was “outside the political reality”… and this new paper was published under the BIS banner…)

Photo: Eagle Business Credit’s LinkedIn page

Some time ago I wrote a blog post highlighting the negative effects of the new banking regulatory framework on financial markets liquidity. Since then, concerns have grown.

Here’s a patchwork of a few of the articles published on the topic over the past few weeks (mostly from the FT).

This article demonstrates the extent of the decline in liquidity in both corporate and Treasury markets:

Here the author points out that the next crisis might have a “paradoxical cause” by forcing financial institutions to purchase the sort of assets that regulators deem safe. The consequence of this being that

corralling a huge amount of capital into a narrow band of the market drives prices to perilous highs. Even if these assets were safe to start with, the enforced concentration is enough to make them risky.

But central banks are making the situation worse with their QE programmes:

At the same time as regulated companies are being forced to buy these “safe” assets, then monetary authorities have been taking away large swaths of supply.

This is spot on. I keep reemphasizing this point too. And the author to brilliantly conclude:

In the popular narrative, the financial crisis was caused by the wilful wrongdoing of the banks. Regulators should know better. In financial markets, risky behaviour is less often born of recklessness than of a false sense of safety.

In this article, Blackrock’s chief investment strategist estimates that, as QE reduces the supply of bonds in the market and replaces it with cash, life insurers and FX reserve funds need an additional $3.5tr more each year to reinvest the proceeds of their maturing bonds and pension funds have an extra $500bn to reinvest every year. Compare this with QE reducing the supply of safe assets, which should decline by $300Bn this year. In the end, the shortage of high-quality liquid assets is estimated at $3.3tr. No wonder a number of assets are trading at negative yields. And this excludes the reinvestment needs of mutual funds (which have $50tr of assets under management) and banks’ regulatory needs…

Artificial shortage of ‘safe’ assets and negative yields go hand in hand.

Despite those facts (mostly coming from practitionners), there are a number of dissenting voices, such as Matt Levine (here on Bloomberg). His reasoning is a bit sloppy. According to him, because dealers would not be buying bonds all the way down during a crash anyway, at least now investors know that there is a risk involved in trading. But this isn’t the issue. The issue here is that we have created more risks to both investors and dealers by amplifying price swings. Investors already know that in a crash, their assets holdings’ valuation is going to decline, dealers dampening the crash to an extent or not. Should we also make cars less safe than they have become through natural manufacturing progress in order to make sure that drivers are careful on the road? While the illusion of safety (or artificial safety) is a bad thing, the artificial amplification of risk also is.

The same sort of logic has been applied by regulators and central bankers (such as Mark Carney) who keep denying that regulation or central banks policies have had anything to do with the current liquidity drought. On the contrary, they warn that it was the pre-crisis period that reflected artificial liquidity!

Same thing here, in an article that completely misses the wider picture. The author retains some of Matt Levine’s arguments (dealers wouldn’t buy all the way down), but also criticises investors for taking too much risk that they “do not understand”. Right. Perhaps making a link to banking regulation, interest rates, QE and shortage of safe assets would allow the author to understand that the search for yield and the artificially high prices of some assets do not come out of nowhere. Investors aren’t that ignorant Mr. Mackintosh. They prove it every week by writing articles warning about those exact risks in your own newspaper. I can send you the links.

Update: Reuters reports that a recent finance industry conference in London pointed out that the situation is going to get worse over the next few years. The audience of the panel reached 200 people, twice as much as for any other panel, highlighting how concerned market actors are.

I have been pretty busy recently so now trying to catch up with a number of things.

I read in The Economist a couple of weeks ago that a new hybrid capital instrument proposal for banks could bypass regulatory decisions and rely on market triggers instead. I decided to investigate.

The principle isn’t very complex at first glance but raises a number of questions. According to Bulow and Klemperer (see their paper here), those equity recourse notes (ERNs) would classify as a type of convertible hybrid capital that avoid the flaws of Basel 3’s contingent convertible (CoCos). The problem with CoCos was already known by Bagehot in the 19^th century: define an artificial threshold and get ready for a panic once you officially announce that this threshold has been breached.

ERNs hope to avoid this problem by getting rid of the regulatory trigger altogether. According to the authors, ERNs are

a form of “contingent capital” that start life as debt, but any currently-due payment is automatically converted into equity if the share price is below a trigger on the day the payment is due. The crucial features are that

(1) conversions are based on whether the share price is below a trigger that is a fixed fraction of the issue-date share price,

(2) conversions into shares value those shares at the trigger price, and

(3) conversions occur one payment at a time–only currently-due payments convert.

Essentially, banks would have the ability to pay interests on ERNs with newly-issued shares rather than cash (thereby safeguarding some precious liquidity in times of stress).

Some extra features of ERNs include:

– ERNs cannot contain covenants limiting the issuance of future ERNs. (See Section 3; there can be covenants limiting the future issuance of conventional debt.)

– Cash dividends and buy-backs are subject to regulatory approval. (This prevents a bank from undermining the role of ERNs by announcing that it will buy back the stock from any conversion or buy back ERNs prior to any repayment that would be converted; of course, major banks already require regulatory approval for dividends and buybacks.)

– Banks have the right to make payments on ERNs in shares (valued at the trigger price) even if the shares are trading above the trigger price. We would not expect to see this option used much, but it ensures that a bank will never fail because it cannot pay off an ERN.

– In bankruptcy, ERNs convert entirely into shares (that is, each ERNholder receives the number of shares equal to the ERN’s face value divided by the trigger price).

– The full benefits of ERNs require restrictions on their term structure.

And to implement ERNs within banks’ capital structure, the authors came up with the following reform programme:

1. Have ERNs replace traditional cocos;

2. Require all SIFIs’ long-term debt be substituted by ERNs;

3. Require secured debt (and defaulted contracts such as loan commitments) to have recourse (beyond specified collateral) only to shares or ERNs;

4. Isolate deposits in well-capitalised subsidiaries, with capital requirements similar to those the private market, or at least central banks, apply when lending against similar assets;

5. Slowly (to allay concerns about, and to respond to, any unintended consequences) narrow the eligible collateral in these subsidiaries.

So ERNs, as devised by Bullow and Klemperer, clearly aren’t some sort of magical free market solution: they still require significant regulatory intervention in order not to undermine the whole mechanism.

The authors also claim that ERNs are counter-cyclical, by facilitating capital issuance in bad times and hence supporting new lending opportunities, and prevent regulatory capital manipulation.

A few points quickly emerge:

It seems like banks could be vulnerable to ‘attacks’ from investors: if a bank’s stock is trading slightly above the trigger, some market participants could well short the stocks, pushing them below the trigger, and then earn a free lunch by benefiting from the effects on the share price of the dilution created by the automatic issuance of further shares. The authors point out that dilution would only be slow. This is true, but if the share price was already trading slightly above the trigger, even a minor dilution could push it permanently below the threshold. The situation would then be self-reinforcing as more investors would sell the stock as they expect further dilution to occur.

In particular in times of stress, when creditors would rather secure cash than extra equity stakes as payment, the automatic selling could push the share price in a downward spiral*. Equally, the mandate of a number of fixed-income investors does not allow them to hold equity investments. Consequently, they would have to get rid of the shares, creating further downward pressure on the share price. The risk here is permanent and self-reinforcing dilution.

The paper claims that ERNs would prevent such downward spiral scenario, as the value of the share issued would be below the value of the cash saved (and that, if ever necessary, banks could anyway buyback shares using the saved cash, which seems to contradict the features they described above, i.e. regulatory approval). While there is indeed some truth in this theory as it would benefit the book value of the bank, the market value is something entirely different and relies on the discounted free cash flows (or dividends as is often the case with banks) that originate from future earnings. Current cash levels do not impact a permanently reduced ability to generate income. Their points also do not address most of the cases I have described above.

A potential solution could be to redeem the ERMs whose trigger price is ‘permanently’ above the share price, although this may not always be possible when contractual or regulatory constraints are present.

The claim that banks will have counter-cyclical incentives to lend is also dubious. In times of stress, leverage can indeed accelerate a bank’s demise. What ERNs do is that banks’ most liquid assets (i.e. reserves/cash) become partly safeguarded, limiting banks’ incentives to call loans or sell securities on the market. But in times of stress, banks are likely to see their cash inflows reduced due to the difficulty of issuing new ERMs on the capital markets, or because the economic situation makes borrowers less creditworthy. It is then highly unlikely that ERNs would allow banks to lend during such periods. At best, they would relief some pressure on banks’ funding structure and liquidity (and leverage to an extent).

All in all, and there may be some critical points that I have missed or misinterpreted (their paper is pretty dense), ERNs aren’t some sort of magical free markets pills unlike what The Economist seemed to imply. However, I have nothing against experimenting: in a world free of regulatory constraints, banks are free to issue the type of securities they wish as long as investors are happy to purchase them. Whether the securities end up doing the work they were initially supposed to or not, markets would learn.

PS: it does seem to me that ERNs could be issued by any sort of corporations. I am unsure why the paper only focuses on banks.

About a year ago, some BoE staff produced a paper on money creation that made some noise by (apparently) embracing an endogenous (outside) money view. A number of people, and I, pointed out the internal inconsistencies of the theory outlined, which in fact seemed to misunderstand what the traditional money multiplier was about.

A year later, here we go again.

A couple of weeks ago, other BoE staff published a working paper explicitly titled Banks are not intermediaries of loanable funds – and why this matters. Surprisingly, not many people seemed to have picked up on it so far. Unfortunately, the same sort of ‘ivory tower’ flaws and contradictions are present in this paper.

The very first page of the report declares that:

The bank therefore creates its own funding, deposits, in the act of lending, in a transaction that involves no intermediation whatsoever.

I just can’t believe that this is a serious assertion. This is repeated throughout the paper. Later they declare that a bank cannot lend more by attracting deposits from competitors (page 10). Given what I keep reading in academic economic research documents, I’d seriously advise all economists to speak to bankers, bank CFOs, bank treasurers, bank analysts and other market participants before writing such things. They would find out that when a bank lends it indeed creates new deposits, although those do not represent funding but extra pressure on the bank’s existing funding structure (and liquidity)!

They would also find out that banks actively compete for deposits, by raising the rate they pay. According to the authors’ theory, bankers are really really confused and should not bother attempting to attract deposits in the first place… This would mean the suppression of the market for deposits, which a quick reality check rapidly dispels. I have spoken to many bankers whose lending and external funding plans were carefully linked (and no, they never considered raising funds from the central bank, as the theory suggests, unless exceptional circumstances).

In fact, the whole purpose of funding is to get hold of extra reserves. Raising deposits (retail, corporate, interbank…) or raising wholesale funding (senior, sub, secured debt, repos) brings about a reserve (high-powered money) transfer from one bank to another. The one that suffers from the deposit outflow sees increased pressure on its funding structure as its liquidity declines. The one that benefits from the deposit inflow, on the other hand, is now awash with new liquidity it is willing to use.

Despite being a BoE paper, it seems like the UK experience has been quickly forgotten: Northern Rock failed exactly because it had funded its growth by raising mostly volatile wholesale funding that vanished all of a sudden when the crisis struck, leaving the bank illiquid and in a downward spiral. This would have never happened if it had been able to ‘create its own funding’.

As I have explained time and time again, it is naïve to believe that banks can simply continuously extend credit then borrow the reserves from the central bank, and get away with it. The central bank lending stigma is strong, and it is confirmed by the literature (see here or here). This fact is well-known by market participants but clearly not by endogenous money theorists, unfortunately. As the paper says:

The [deposit multiplier] model however does not recognise that modern central banks target interest rates, and are committed to supplying as many reserves (and cash) as banks demand at that rate. The quantity of reserves is therefore a consequence, not a cause, of lending and money creation.

It’s not that central banks don’t provide the reserves. It is that commercial banks are not willing to borrow them!

Ironically, the paper contradicts itself when it says (page 5, my emphasis):

They add that, from the (microeconomic) point of view of an individual bank that considers whether to deviate significantly from the behaviour of its competitors, other important limits exist, especially increased credit risk when lending too fast to marginal borrowers, and increased liquidity risk when creating deposits so fast that too many of them are lost to competitors.

Another problem of this paper is that it wrongly depicts the traditional money multiplier model. Banks can lend before they secure extra-reserves, as long as they have enough reserves to face adverse clearing during the interbank settlement process, or when the new loan is withdrawn as cash. See the following extracts as evidence they do not understand that the money multiplier model does allow for endogenous money creation through deposit creation (page 6):

The fact that banks create their own funds through lending is acknowledged in descriptions of the money creation process by leading central banks and policymaking authorities. The oldest goes back to Graham Towers (1939), the then governor of the central bank of Canada: “Each and every time a bank makes a loan, new bank credit is created — new deposits — brand new money”.

And (page 7):

The fact that banks create their own funds through lending is also repeatedly emphasised in the older economics literature. One of the earliest statements is due to Wicksell (1906): “The lending operations of the bank will consist rather in its entering in its books a fictitious deposit equal to the amount of the loan…”. Rogers (1929): “… a large proportion of … [deposits] under certain circumstances may be manufactured out of whole cloth by the banking institutions themselves.”

And (page 13):

The fact that the creation of broad monetary aggregates by banks comes prior to and in fact may (if commercial banks need more reserves) cause the creation of narrow monetary aggregates by the central bank is acknowledged in many descriptions of the money creation process by central banks and other policymaking authorities. The oldest and clearest comes from Alan Holmes (1969), who at the time was vice president of the New York Federal Reserve: “In the real world, banks extend credit, creating deposits in the process, and look for the reserves later.” This is exactly the view put forward in this paper.

The authors are mixing up ‘funding creation’ and what Chester Arthur Philipps, in his 1920 book Bank Credit, called ‘derivative deposits’, or temporary deposits created by the bank by the action of lending:

In fact, nobody has ever denied the endogenous money creation inherent to the money multiplier model. But this endogenous creation of deposit liabilities (inside money) is constrained by the exogenous variable of the availability of reserves (outside money), as I explained in this post. Is there a fixed limit in the absence of reserve requirements? No. But in this case banks estimate the amount of precautionary reserves and secondary reserves (i.e. mostly highly-rated/high-quality liquid securities they invest in for margin and liquidity management) they need. Apart from asset quality considerations (and other exogenously-defined factors like banking regulations), nothing prevents a bank from expanding its loan book, and hence its liabilities, ad infinitum. Except the threat of illiquidity.

During the Scottish free banking period, it was reported that banks maintained minimal reserve ratios of around 2% once clearinghouse were established, but sometimes as high as 60%+ before that (see Selgin’s The Theory of Free Banking). Things aren’t that different today: markets are more liquid, securities types more varied, volumes higher, facilitating day to day liquidity management, but the central bank funding stigma prevents banks from having a liquidity free lunch.

So in short, in the absence of reserve requirements, what we end up with is a banking system whose endogenous expansion is bound by the (also inherently endogenous) market actors’ perception of the need for an exogenous variable (reserves/high-powered money). (Not easy to formulate…)

The authors also don’t address the fact that many central banks (China, Turkey…) actively, and rather successfully, use reserve requirements as a monetary policy tool (the latest one being Moldova, but also see examples here).

Funnily, this new BoE staff paper on endogenous money theory disagrees with… last year BoE staff paper on the same topic. While last year’s paper considered that James Tobin’s famous publication Commercial Banks as Creators of Money fitted into their theoretical framework (by describing the limits to bank expansion), this year’s paper mostly views it as in contradiction with their own framework. This makes the whole endogenous outside money theory even more confusing.

Even more problems appear when they devise their model (page 12). Their banking system consists of a single bank that can extend credit without limitations. They fall in the trap that I explained in my detailed critique of the MMT version of the endogenous money theory (see also here). With no competitor expanding or contracting at different paces (as in the real world), this single bank is indeed freed from adverse clearing constraints on liquidity (with only cash withdrawal risk remaining, although cash is very rarely mentions in this article). And worse, they use Werner’s rather weird research paper (in which he ‘discovered’, for ‘the first time in 5000 years’, that customer deposits are part of the bank’s balance sheet… Yep, really…) as ‘evidence’ empirically validating their view (see my more comprehensive review of Werner’s piece here).

Let just finish with the fact that this latest BoE staff working paper never ever mentions any critique of endogenous outside money-type theories, and believe that DSGE models are appropriate for macro-economic modelling despite more and more people now turning against them.

In the end, this paper is a ‘rougher’ version of the theory than the piece published a year ago: fewer internal inconsistencies but also less realistic. Seems like the endogenous theory needs refinement.

PS: The authors say that S&P, the rating agency, endorsed their view. This is wrong. The paper they are referring to is a personal report published by the member of one of S&P’s economics department. S&P explicitly says that this report reflects the view of this economist only. He probably should have consulted his bank analysts before publication.