The rights and obligations of principal and agent need to be properly defined, particularly in complex business like banking and finance. Five banks have recently been fined $5.5 billion over a rate manipulation scheme that has seen them not act in their clients’ best interest.

A bank should be clear about whether it trades as principal or agent when it transacts with a client. If as principal, the rules of disclosure may be relaxed. If as agent the rules of disclosure are clear: the client must be made aware of the detailed economics of the trade including the commissions, costs and expenses. The concept of markup pricing is incompatible with an agency trade. In fact, not only the quantum but the beneficial recipients of commissions, costs and expenses should be transparent, so that there is no ambiguity as to the interests of the agents and their delegates or associates. For principal trades, the client needs only know the all in cost of the transaction. Margins and markups, and their beneficial recipients are irrelevant.

This transforms the issue from one of transparency of pricing to one of the distinction and separation of principal or agent relationships. Client’s may want to choose whether the bank they trade with is trading as principal or agent. If there is no liquidity, it may be preferable to do a principal trade since the bank makes market. If there is ample liquidity, an agency trade may be preferable as pricing is transparent. What is required is a Chinese Wall between the principal desk and the agency desk. If a client chooses to call the agency desk, they receive full transparency but have to live with the liquidity available. If they choose the principal desk, they are aware that the bank is trading as principal and does not in any way guarantee best execution but the bank must guarantee execution.

Incidentally, the complaint against the banks was not that they did not act in the best interest of clients, but that they colluded to create a false market, or lack thereof, and distort prices. In a fair market, even if all transactions were principal ones, clients would have obtained price discovery by shopping around.

The current convention is one where banks trade as principal and therefore, rightly should have no obligation to provide transparency of pricing or best execution. When trading as principal the bank acts in its own best interests, not that of its so-called clients. The clients of the bank, when entering a principal trade become counterparties for the purposes and duration of the trade, and counterparties are owed a different set of obligations than clients or customers. The possible source of confusion is that clients are unaware or unaware of the implications of being in a principal trade. They may be under the impression that the bank acts in any way in their interests, a clearly mistaken assumption. The fault of the banks, if any, is to perpetuate the myth, actively or passively, that they in fact act in clients’ best interests. Where there is a fiduciary responsibility to do so, the law compels them to act in the best interests of their clients but where there is no such relationship, clients should beware.

Regulators can clear the situation by distinguishing between principal and agent transactions and setting out the standards of behavior in each relationship.

It would certainly be interesting to see, in a free market, which business, principal or agency, finds more custom and which is more vigorously supplied. Thin and thinning agency margins balance regulatory capital requirements needs to support principal businesses and only an unfettered market providing both alternatives will complete the market for these services. It is likely that with clarity and the clear distinction between business lines, new entrants and innovation will lead to more efficient markets less prone to abuse.

On Monday, 18 May, Jiangsu province successfully sold 52 billion RMB of municipal bonds.

Recently, the Chinese government has accelerated a municipal bond program that effectively is a debt swap for some 1.7 trillion RMB of existing local government debt. Cost of debt for the local governments is expected to fall by some 250-300 basis points.

There is some logic behind what is effectively a system wide restructuring. The municipal bonds yield less than the maturing or retired debt, however, they also have a lower risk weight of 20% or less and thus consume a lot less capital.

For a bank seeking to maintain profitability, it would have to extend more loans for every dollar notional of restructured debt. The municipal bonds are financed at highly advantageous rates and haircuts. Naively, if risk weights were reduced to 20%, a bank could make 5 times the loans it used to with the same capital and fund itself cheaply, not with deposits where reform is raising rates, but with medium to long term repos with the PBOC, namely the MLF and PSL. Impact on bank profitability would be neutral, balance sheets would expand, loan growth would rise, credit quality would improve, and the impact will likely be felt across the economy as a whole and drive the stock market even further.

At the same time, the Chinese regulators have relaxed funding conditions by allowing banks to refinance existing LGFVs in respect of projects started prior to the end of 2014. This is likely a stop gap measure as the PSLs are unlikely to be able to absorb the volume of refinancings of what is essentially the substantial majority of local government debt in the system.

China is initiating QE lite mimicking the ECB’s LTRO programs. The objectives are clearly to lower borrowing costs for local governments, and to establish a liquid municipal bond market as part of a reorganization of the funding mechanism for local governments which until now had used opaque, off balance sheet, so-called Local Government Funding Vehicles (LGFVs). The opacity of LGFVs lead to uncertain priority of claim and difficulty in measuring systemic risks. The side effects of QE Lite will include expanding liquidity and bank lending through the eligibility of municipal bonds as collateral in PBOC repo operations.

The PBOC has effectively halted the creation of new LGFVs, is encouraging local governments and their creditors, mostly commercial banks, to exchange LGFV debt for municipal bonds. It is also encouraging private commercial banks to invest in municipal bonds which are eligible for repo and thus provide a nexus between the PBOC and local governments.

Initial efforts have been clumsy due to the inexperience of the parties, the banks, the local governments and the PBOC. For the strategy to work, the banks have to be de facto guaranteed a return on capital exceeding borrowing costs for buying the muni bonds.

1. The bonds have to offer sufficient yield over the repo rate and other competing investments consuming the same amount of capital. This would include treasuries.

2. The capital consumption, or the risk weight assigned to these municipal bonds has to be sufficiently low so that, returns notwithstanding, they consume little or no capital.

3. The collateral standards have to be sufficiently accommodative.

The above conditions have to be in place before any muni issuance takes place. The PBOC may need to bootstrap the process by communicating and educating the banks.

Next week, on May 18, the Jiangsu government will issue 52 billion RMB of bonds, 30.8 billion of which are refinancings. The bond issue was initially planned for Apr 23 but was postponed for lack of interest. We wish the Jiangsu government and the PBOC a healthy bid to cover.

Why did German bunds selloff so violently?

When the ECB announced QE, traders attempted to front run the program. The program would buy bonds according to the capital subscription of the national central banks to the ECB. On that count, bunds would see an 18% allocation of the budget, France 14%, Italy 12% and Spain 9%. Traders reasoned that bunds would see the biggest allocation not only in absolute terms but relative to the stock of bonds available. The problem with this thesis was the investors who had over 2 years of profits ready to crystallize.

A less volatile trade expression for QE was to realize that with risk mutualisation, 20% explicitly and 100% implicitly through TARGET2, it made sense to sell protection on Italy, Spain, Portugal and to buy protection on Germany and France.

When the Maastricht treaty was signed, peripheral spreads over bunds were much wider than they are today, some 4%-5%. They spent the best part of a decade converging, in the case of Italy, to a negative spread to bunds… In 2012, spreads widened to pre Maastricht levels as country risk reasserted itself in the European sovereign bond market. Since the ECB’s “all it takes” policy, spreads have converged again. We are currently still well wide of zero and convergence remains a logical trade.

The Chinese economy generated about 10.4 trillion USD of nominal output in 2014, representing incremental nominal output of 680 billion USD, equivalent to 7.0% growth. If the Chinese economy continues to generate this same incremental nominal output of 679 billion USD this year and next, 2015 growth will be 6.54% and 2016 growth will be 6.14%. This naïve calculation provides one potential lower bound estimate for a safe landing for the Chinese economy, that is growth rates that would not trigger significant unemployment, deflation and social unrest. On this trajectory, China will have slowed to mature market rates of growth by 2030.

Incidentally, the Economist forecasts growth of 6.90% in 2015 and 6.80% in 2016.