PrefBlog

BIS has released a working paper by Ben Craig and Goetz von Peter titled Interbank tiering and money center banks:

This paper provides evidence that interbank markets are tiered rather than flat, in the sense that most banks do not lend to each other directly but through money center banks acting as intermediaries. We capture the concept of tiering by developing a core-periphery model, and devise a procedure for fitting the model to real-world networks. Using Bundesbank data on bilateral interbank exposures among 1800 banks, we find strong evidence of tiering in the German banking system. Econometrically, bank-specific features, such as balance sheet size, predict how banks position themselves in the interbank market. This link provides a promising avenue for understanding the formation of financial networks.
…
The first column of Table 1 reports the simplest regression using bank size as the sole explanatory variable. The log of total bank assets is highly significant; a marginal increase in size from the average balance sheet of &eur;230 million raises the probability of belonging to the core by a sixth of a percent. Indeed, size is a fairly reliable classifier. The average size of banks in the core is 51 times that of banks in the periphery. Hence, large banks tend to be in the core, while small banks are found in the periphery of the interbank network.
…
Systemic importance is highly correlated with a bank’s network position: it is extremely unlikely that a systemically important bank would not be in the core, as indicated by the low rate of false core predictions, Prob(c|P). But the moderate fit also suggests that a bank’s position in the network is something that goes beyond its systemic importance.

The success of “Balance Sheet Size” in predicting network placement supports my contention that surcharges for systemically important banks should be based on a balance sheet variable – either gross assets or risk-weighted assets, I’m not particular – rather than upon a regulator’s determination that such-and-such bank is systemically important.

One thing I would like to see addressed is an examination of how increasing the risk-weighting of interbank assets would change the model. It is far too cheap, in terms of capital, for a bank to hold another bank’s paper; this increases the chance for contagion.

It was a day of readjusting prices on continued heavy volume today, with PerpetualDiscounts losing 8bp and FixedResets picking up 8bp.

Japan is rumoured to be considering a capital surcharge on systemically important banks, but details and confirmations are scarce:

Japan’s financial regulator is considering forcing the country’s largest banks to hold more capital than required under Basel III rules, a person with direct knowledge of the matter said.

The Financial Services Agency will start internal discussions soon on whether to apply a capital surcharge to systemically important lenders such as Mitsubishi UFJ Financial Group Inc., the person said, declining to be identified because the matter is confidential.
…
The Swiss panel said UBS AG and Credit Suisse should hold almost double the capital required under the Basel III proposals announced last month. By 2019, the lenders would need to hold at least 10 percent of capital in common equity, compared with 7 percent required under Basel.

The possibility of a global capital surcharge of around 2 percent for the world’s most important banks “cannot be ruled out,” Shinichi Ina, a Tokyo-based analyst at Credit Suisse, wrote in a report this week.

This is regulation in a nut’s hell. I support a progressive surcharge on Risk Weighted Assets. I am quite aware that RWA is, at best, an imperfect measure of a bank’s systemic importance; but I assert that singling out a group of “Top Tier” banks is worse.

Bad news for Ireland:

Ireland got an unwelcome jolt Wednesday as Fitch Ratings cut its sovereign credit rating to the lowest level of any of the major agencies, citing the heavy burden of bailing out the country’s banking sector.

Fitch cut Ireland’s long-term foreign- and local-currency ratings to A+ from AA-. Moody’s Investors Service, which warned Tuesday that it was reviewing the country for a possible downgrade, and Standard & Poor’s both rate Irish debt higher.

The move “reflects the exceptional and greater-than-expected fiscal cost associated with the government’s recapitalization of the Irish banks, especially Anglo Irish Bank,” Chris Pryce, Fitch’s director of sovereign ratings said in a statement.

The Bank of Canada has released a Working Paper by James Chapman, Jonathan Chiu, and Miguel Molico titled Central Bank Haircut Policy:

We present a model of central bank collateralized lending to study the optimal choice of the haircut policy. We show that a lending facility provides a bundle of two types of insurance: insurance against liquidity risk as well as insurance against downside risk of the collateral. Setting a haircut therefore involves balancing the trade-off between relaxing the liquidity constraints of agents on one hand, and increasing potential inflation risk and distorting the portfolio choices of agents on the other. We argue that the optimal haircut is higher when the central bank is unable to lend exclusively to agents who actually need liquidity. Finally, for an unexpected drop in the haircut, the central bank can be more aggressive than when setting a permanent level of the haircut.

The TMX has a new page showing Bond Market Data. The data from today (labelled July 19, 2010), showing $20.04-billion face nominal Canadas being traded vs. $0.01-billion face RRBs, should illustrate the points I have made at various time regarding the relative liquidity of the latter! I presume – but do not know – that they get their trading statistics from the same data set used for their prices, which are major Canadian dealers.

The Canadian preferred share market showed continued strength on continued high volume, with PerpetualDiscounts gaining 17bp and FixedResets up 10bp.

PerpetualDiscounts now yield 5.47%, equivalent to 7.66% interest at the standard conversion factor of 1.4x. Long corporates now yield 5.2% so the pre-tax interest-equivalent spread (also called the Seniority Spread) now stands at about 245bp, a sharp decline from the 260bp reported on September 30.

The European stress tests are encountering renewed criticism:

The health check, undertaken by the Committee of EuropeanBankingSupervisors (CEBS), tested sovereign debt holdings in short-term trading books but turned a blind eye to the longer-term banking books. Then it emerged that the so-called “gross disclosures” weren’t gross at all. The latest Irish bank bailout has now revealed a third flaw.

The tests were supposed to reassure global investors by showing that systemically important eurozone banks could survive what was perceived as their biggest threat — a sovereign default. But the tests skirted around the Irish situation in two ways. First, Ireland’s banks have failed due to an immense property crash, not a sovereign crisis. Second, the biggest disaster zone — Anglo Irish Bank — was not even stress-tested.

Jerome Kerviel, the SocGen trader who was in the position of a drug mule caught at customs, has been sentenced to three years in jail:

Former Societe Generale (SOGN.PA) trader Jerome Kerviel was sentenced to three years in jail by a Paris court on Tuesday for his role in a trading scandal and ordered to pay the French bank 4.9 billion euros ($6.8 billion).

The verdict came as a victory for SocGen, which always maintained Kerviel acted alone and without the sanction of his managers at the bank. It had sought payment of damages for the money it lost unwinding the trader’s risky market bets in 2008.
…
The payment to SocGen equates to 3.2 percent of France’s central government deficit for 2010, the GDP of Monaco or 16 percent of the French bank’s market value. Kerviel is currently paid 2,300 euros a month as a technology consultant.

Kerviel was last discussed on PrefBlog on May 3. There is no indication that any of those actually responsible have even been charged.

Moody’s warns on Ireland:

Moody’s Investors Service said Tuesday it may cut Ireland’s debt rating again, citing the increased cost to the government of repairing the stricken banking system, weak economic growth and rising borrowing costs.

A further downgrade could push Ireland’s borrowing costs higher and make it more difficult for the government to meet its debt repayments without seeking help from the European Union’s European Financial Stability Fund.

“Me, too!”, shouts DBRS. The EU bureaucrats will have to take decisive action, as discussed October 1.

Norwegian police are feverishly attempting to make the markets safe for the incompetent:

Two day traders have been arrested by Norwegian police for allegedly cracking an algorithm of U.S. brokerage firm Timber Hill, in order to manipulate the stock prices of three companies listed on the Oslo Stock Exchange.

The two day traders, Svend Egil Larsen and Peder Veiby, have been charged with “market manipulation” and face up to six years in jail if convicted. Timber Hill is a subsidiary of Interactive Brokers and acts as its market maker.
…
Norwegian police claim that between March 2007 and March 2008, the two traders conducted more than 2,200 purchase and sale orders which were “ not real” and jacked the price of the shares in three Norwegian companies up or down before taking a profit Bottom line: the two day traders outsmarted Timber Hill’s algorithm and made money in the process. The three Norwegian companies were Hafslund, Wilh. Wilhelmsen and Odfjell.

Veiby earned 250,000 Norwegian kroner ($40,698) in the alleged scam while Larsen earned 160,000 Norwegian kroner ($26,056). Timber Hill did not return Securities Technology Monitor’s call seeking comment by press time.
…
“In our view it is a deliberate manipulation against the computer they’ve been trading against so that the system changed the prices and they were able to earn money,” said Christian Steinberg with the Norwegian National Authority for Investigation and Prosecution of Economic and Environmental Crime in an interview with the Norwegian newspaper Dagens Naeringsliv.

The newpaper report said that Larsen declined to comment and Veiby denied any wrongdoing on the grounds he did “not act with the intent to commit price manipulation in the legal sense.” The alleged scam was uncovered by the Oslo Stock Exchange, which reported it to the Norwegian National Authority and Prosecution of Economic and Environmental Crime.

I don’t see that anything wrong was done at all. Willing buyer, willing seller, arm’s length … what’s the problem? Trades “not real”? I don’t see it. They owned the stock, they had the risk, there was transfer of risk … if Timber Hill can’t programme its algorithms better than that, the individuals charged have done a service to the market by taking capital away from those unable to deploy it with competence. Timber Hill’s only possible excuse is that the amounts were so small (less than $100,000 total profit for a year) that the efect was lost as a rounding error … but that’s a pretty flimsy excuse.

However, according to the Google translation of a story in Norwegian titled Klart dette er greit, the story is more complex than reported above – naturally enough:

The two exploited a weakness in the computer algorithm in the near half a year without Hill Timberlake discovered it. This week in Oslo District Court has been discussed in court if the two ran the illegal market manipulation, or whether they simply were smarter than any other market player.

– It is obvious that what they have done is okay, says Jan Erik Meidell.

Meidell was employee number three when he started in Timber Hill in 1994. He helped build the brokerage house’s position in market making in Europe, and was one of two managers for their trading in Europe. Meidell ruled the so-called share the robots, and he claims to know the Timber Hills algorithms and trading strategies in detail. Today Meidell PhD student at NHH and co-owner of the investment fund NorthSeaGem.

He believes it is “nonsense” that the two persons now sitting accused of having exploited a weakness in Timber Hills algorithms.

– Timber Hill has chosen a strategy that involves risk and acknowledge that it sometimes will lead to losses, “says Meidell to dn.no
…
– What do you think that people take advantage of the robot’s weakness when the opportunity occasionally arises?

– Sure it’s okay. Whoever is to play, must withstand the roast, “says Meidell.

The most important point is that Timber Hill is fully aware that they can not have full control over all shares always mean Meidell. It is not so important – the overall strategy is in fact profitable strategy for Timber Hill.

– They think that what they lose in one stock, it will win them back in others. Does the strategy in more than 50 percent of cases, so you earn bucks, and it’s just that Timber Hill, says Meidell.
…
Timber Hill has been absent in the trial which is now in the Oslo City Court, and they have not wanted to respond to Dagens Næringsliv inquiries.

– I think Timber Hills reaction ultimately had been “awesome, you managed to crack our algorithm. Perhaps they had even offered them jobs, “says Meidell.

So .. who knows? However, it does look a little as if the Norwegian regulator is attempting to ensure that incompetent traders are protected, and that the stock market is a nice little place where children play nicely.

Another strong day on heavy volume for the Canadian preferred share market, with PerpetualDiscounts up 30bp and FixedResets tagging along for a 9bp gain.