Letter from Professor Spahn to the Committee
I have the pleasure of responding to your questions
1. One question which was asked to Professor
Spahn but not answered at any length was "what makes you
think that the imposition of a tax wouldn't lead the FOREX markets
to migrate in the same way as dollar banking moved to Europe and
generated the Eurodollar markets in the 1960s/70s"?
Answer: There are many factors that explain
the emergence of the Eurodollar market in the 1960s/1970s:
First, there were unremunerated minimum
reserve requirements on bank deposits to be held at the Fed.
This imposed a charge on US dollar holdings in bank accounts within
the US. Elsewhere it would not apply, so there was a trend to
migrate to other places in Europe (and Asia). The cost advantage
is not only confined to injections of primary liquidity into the
Eurodollar market (US dollars flowing to Europe), it also applies
to all credit operations that are based on this primary
liquidity (and the credit-multiplier was very high in this
The same would of course also apply
to holdings of DM in Germany, Pound sterling in the UK, Francs
in France, etc, so there was a trend to leave domestic currency
markets all over the world. It led to a concentration of foreign
currencies in London, mainly because it was (and is) the largest
financial centre that benefits from a number of structural and
institutional advantages. In particular it had (and has) the deepest
and most liquid wholesale market in the world, where both creditors
and borrowers can benefit from substantial economies of scale.
Other reasons for the dollar to go
overseas were related to structural deficiencies of the US financial
markets (now partly removed). There was a cap on the interest
rate for savings deposits (regulation Q), but this could largely
be circumvented by other instruments (eg, money market funds)
even in the US. There seems to be agreement that the temporary
interest-rate equalisation tax of the 1960s was largely ineffectual.
Other obstacles were more difficult to deal with, in particular
the regional segmentation of the capital market. (As late as 1979
there was legislation pending in the US Congress to forbid national
non-bank holdings that were organising supra-regional liquidity
Finally, in the early 1970s, the
depositors of petrodollars preferred London to New York for convenience
and political aspects. Some oil producers feared political coercion
by the United States, even sequestration of their assets, a concern
that was shared by the Soviet Union and other socialist countries
that would also prefer London for their dollar holdings.
This set of conditions (only partly financial)
cannot be compared with a situation where a 0.01 percent charge
would be levied on transactions (which is of course negligible
for the long-term investor who was then much harder hit by the
implicit charge of reserve requirements).
2. It was mentioned that a recent proposal
sees the tax as being a percentage of the bid-ask spread, rather
than a percentage of the total transaction. Please could you provide
the Committee with a short clear explanation of this? Also, is
this a Schmidt-inspired proposal?
Answer: I am afraid the answer is not short.
I may have been contributing to this idea myself because in hearings
last year I wanted to draw the attention to the fact that bid/ask
spreads differ for various currency pairs. And I had argued that
a transactions tax on the bid/ask could be interpreted as a presumptive
gross income tax (in order to convince those that are against
the CTT, but raise concern about the loss of sovereignty in taxing
the returns on international capital investments under globalisation).
I have revised this position mainly on the following grounds:
While the bid/ask is known at any
one moment, it cannot be applied to one single transaction, because
the conditions for traders may have been different when taking
an open position from those at the time of closing the position.
The profit margin may be greater than the bid/ask, it could even
The margin varies during the day,
it is different among different types of traders, and even for
individual clients. The bid/ask for liquidity traders is normally
much lower than for dealings with non-financial institutions.
There is no record on the margin
by trade (even if it were, I am horrified by the reporting requirements).
The only alternative is then to tax the accumulated margins at
the end of the business day. It would come close to taxing gross
income on a daily basis. This would not represent a transactions
tax, so the benefits of driving out smaller noise traders would
be lost. Even if done once a day, it would be too much reporting
and too much intrusion into the banking business. And, of course,
a tax rate on the accumulated margins would have to be much higher
than any transactions tax rate (if the top income tax rate were
30 per cent, the tax rate on the accumulated margins as an indicator
of gross profits could be as high as 20-25 per cent). This would
work as a huge deterrent.
Finally, since the margin is known
neither to the trader at the desk, nor at the point of settlement,
the tax would have to be levied at the back office, the only place
to possess the information. But most actors agree that a back
office (unlike the trading desk and/or the clearing/settlement
operator) is much more difficult to control, and it could indeed
be more easily shifted out of the jurisdiction.
There is also concern that a CTT would hit emerging
and developing countries more heavily, and therefore some propose
to exempt this market altogether. One has to realize however that
these smaller and much less liquid markets operate with enormous
bid/asks (often in the order of percentage points rather than
basis points). If a tax were put on the margins of these tradings,
it would particularly penalise the Third World. (The reporting
requirements would of course abort such an attempt. Fortunately!
Because I am strictly against too much reporting and certification
in developing countries, since it encourages corruption!)
3. Finally, there's a question about
the width of the corridor in the two-tier variant, and the rate
of change allowed in the target and hence the corridor (which
also determines the sorts of speculation which is deemed excessive).
Is it your feeling that such issues are political rather than
Answer: Since I assume this tax to be implemented
unilaterally, the political implementation is easier than for
a coordinated approach. A parliament (say, that of Brazil) could
determine the anchor currency (US dollar, or a mix of US dollars
and other currencies), it would determine the period for the moving
average and its weights to calculate tomorrow's target rate, and
it would fix the corridor. While this can be viewed to represent
a technical problem, it has of course political implications.
The basket of currencies for anchoring
is decisive for the direction of trade and capital market policies
(which is a political aspects).
The moving average may be seen to
represent a purely technical matter. But the length of the moving
average and the weights (one could put equal weight on the daily
information, or weigh more recent information more strongly) will
be decisive for the time profile of the longer-term devaluation/revaluation
of the exchange rate (a political question). The longer the moving
average and the heavier the weights placed on historical information,
the closer the scheme comes to a fixed-exchange rate system; the
shorter the period of adaption and the heavier the weights on
more recent information, the faster can the target rate incorporate
new information about fundamentals, and react correspondingly.
I think something like 20-30 business days could be appropriate.
Also the corridor is a matter of
practicality. I suggest to make an empirical study on the daily
fluctuation of a currency during phases of normal trading (which
is a technical matter), but allow some margin of tolerance (say,
twice the normal variability) to avoid the tax to be triggered
too often (a political decision).
Thank you for your challenging questions. I
hope to have served your purposes.
4 May 2002
18 As an aside: the implicit charge is proportional
to the maturity of the deposit, so it penalises long-term holdings
more than short-term commitments, which runs counter to the idea
of Tobin. Back
One could also imagine the conditions to be phased in. In Argentina,
for instance, I proposed (in 1996!) to move from the fixed-exchange
rate currency board system to a more flexible peg be starting
with long periods for the moving average, and shorten it over
time to render it more flexible. Back