Supplementary Written evidence submitted
by Professor Charles Goodhart
At the Treasury Committee hearing the Chairman
asked me to expand on two issues. These are whether either Basel
I or II ever achieved anything positive, and my views on CoCos,
Conditional Contingent Capital, and Bail-in proposals. If I may,
I will add a third brief comment on the coordination problem of
reforming remuneration.
BASEL I AND
II
The main purpose of Basel I was to reduce the,
previously quite rapid, drift down in bank capital ratios. In
this it succeeded, and a time-series chart of such ratios show
that the downwards trend halted in 1988-91, (the transition period).
This success, however, was partially, perhaps largely, later vitiated
by banks' ability to replace core equity by various kinds of hybrid,
and less loss-absorbing, capital instruments, so that core equity,
by 2007, was often in many banks not more than 2% of total assets.
The main purpose of Basel II was to make regulatory
capital more sensitive to risk. While it had some success, it
failed overall both because risk was poorly assessed by both bankers
and regulators, and because risk always appears to decline in
booms (asset bubbles) and to increase in busts. In conjunction
with the greater use of "mark-to-market" accounting,
this made the whole regulatory system far more procyclical, almost
a "Doomsday machine".
COCOS
AND BAIL-INS
FOR FAILING
BANKS
I attach my article (Central Banking, August
2010) on CoCos.
The "bail-in" proposal is somewhat
similar. Assume that Bank A has a loss, too big for its equity
shareholders to absorb. The idea is that the regulator should
then, starting with the most junior creditor, forcibly require
such (subordinated) bonds be transformed into equity, until the
bank is sufficiently well endowed with new equity to carry on
as a continuing entity. There is no bankruptcy, nor cost to taxpayers.
The previous shareholders are diluted into oblivion, and the "hit"
to bondholders varies according to the scale of loss, but could
be large, especially for the most junior.
If only one bank was in trouble (an idiosyncratic
shock), this would be wonderful, the way to go. The problem is
that, with a systemic shock, all banks are, more or less, in difficulties.
The shock to the bond holders of the failing bank would be transmitted
to the bond market and bond prices of all other banks. No bank
would then be able to raise new bond (or equity) finance to recapitalise
themselves. Thus in the recent crisis, to facilitate raising new
long term capital, governments have felt forced to guarantee all
bonds.
If a bail-in system was to be used, whenever
a crisis struck, governments would have to accompany losses to
existing creditors with guarantees for "new-money" creditors,
resulting in a complex mixture of both guaranteed and non-guaranteed
debt. Such non-guaranteed debt would be subject to equity-style
downside risk with no upside countervailing option value. So if
we were to move explicitly to a bail-in system, the cost, or interest
rate, that would be required on such non-guaranteed debt would,
almost certainly, rise very sharply.
Overall it would just seem simpler, and just
as efficacious, to require much higher holdings of equity capital.
REFORMING REMUNERATION
There is a major coordination problem. One bank,
trying to cut back on bonuses, etc, would lose all their top traders
to other banks, as D Maughan found when he tried this at Salomons.
Similarly one country trying to do so would
lose all its top banks/bankers to other countries.
It would be extremely difficult to get effective
international agreement on this subject, partly because it is
so politically sensitive, with different countries having differing
positions, and partly because being the stand-out bank/country
paying more than the rest would be so advantageous.
So nothing can/will be done.
15 September 2010
|
