APPENDIX 3
Memorandum by the British American Security
Information Council (BASIC)
1. SUMMARY
1.1 This is a paper that has benefited from
extensive discussions with economists both inside and outside
government, and reflects part of our joint ongoing work to value
subsidies to arms exports with Oxford Research Group and Saferworld.
ECGD measures its subsidy using a Value at Risk model, which while
conceptually plausible, is full of uncertainties, rendering its
results highly dubious.
1.2 ECGD provides insurance credit guarantees
to exporters, a service available on the open money markets. Our
approach picks up on the government's policy, outlined in its
public-private partnerships and private finance initiatives, to
use the highly-developed finance markets to provide financial
services. If it is appropriate to involve the private sector in
financing public investment such as schools, hospitals and the
tube, why is it appropriate for the public sector to be engaged
in guaranteeing private sector (arms trading) operations?
1.3 The answer to this question frequently
given is that the UK government is in a better position to accept
the risk implied in supplying guarantees than the private money
markets. But if government officials, through privileged information
or through risk reputation, were able to provide better terms
at no extra cost, then there would be a strong case for HMT to
engage in general bond market or credit-derivative speculation.
This clearly is not an option, and would run counter to macro-economic
theory, not to mention every government.
1.4 The money markets are therefore as good
a pricing mechanism as government officials (arguably a great
deal better!). We would therefore argue that the best measure
for ECGD subsidies is to compare the premium rates offered by
ECGD with those obtainable from financial markets through tools
such as government bonds and credit derivatives. We have not got
access to such priviledged information from ECGD, but have been
able to estimate the scale of the subsidy to defence exports,
and believe it to be in the order of £150 million annually.
1.5 The beneficiaries include the overseas
recipient, the UK exporter, and the UK Bank that receives an interest
rate on their loan with no risk at all. But many of these are
transactions happening outside the envelope of economic benefit
(when taking account the opportunity cost). However, taken as
a whole, this is an activity does not benefit the UK economy as
resources are allocated inefficiently. The costs are born by the
UK government (in a reduced credit rating), as well as in the
indirect consequences of the greater trade in arms that results.
BASIC
BASIC is an independent research organisation
that analyses government policies and promotes public awareness
of defence, disarmament, military strategy and nuclear policies
in order to foster informed debate. BASIC has offices in London
and in Washington and its Council includes former US Ambassadors,
academics and politicians. Further information is available on
our website, www.basicint.org
2. ECGD'S ACTIVITIES
2.1 Put simply, ECGD insures exposures to
sovereign governments. Defence companies can buy insurance at
a discounted premium from ECGD against non-payment by customers.
We go into some detail to describe the argument as it is complex,
and the subsidy amounts to more than half the total subsidy we
use in this study.
2.2 The latest figures in the ECGD's Annual
Report show a total support granted in 2002-03 of £3,532
million, exactly half of which was defence-related.[1]
3. DIFFICULTIES
WITH THE
TRADITIONAL METHOD
OF VALUING
ECGD SUBSIDIES
3.1 Previous studies, such as those conducted
by ECGD and NERA[2],
have used a value at risk model to estimate the government subsidy
to the UK arms exports from ECGD. The economists (MoD and independent)
who wrote the November 2001 "York Report" on the economic
impact of a halving of defence exports, also used a value at risk
model to estimate the subsidy to be between £16 million and
£96 million a year. We have no conceptual difficulty with
the "return on capital requirement" methodology used
by the authors. However, this approach suffers from considerable
uncertainties that we believe render the end result highly unreliable.
These uncertainties relate to the amount of required capital and
the required return thereon. We discuss each in turn below.
3.2 First, the value at risk (VAR) methodology
used to calculate the capital requirement is predicated on an
estimate of future volatility. This in turn is based upon prior
historic volatility. ECGD is essentially exposed to sovereign,
currency and interest rate risks over long contractual periods.
We have considerable doubt as to whether historic volatility can
be used to infer very much about future volatility in the sovereign
risk market, even at the level of a portfolio of such exposures.
Any such estimates are subject to such considerable uncertainty
as to render them almost useless.[3]
3.3 Second, the return on this required
capital is estimated by ECGD and the authors of the York Report
at around 5%. This figure, whilst supplied by the ECGD, is not
justified. In the case of defence-related business it also appears
to contradict the Government's own evidence given to parliament
in 2000, which implies an average annual historical cash loss
on such contracts of £52 million over the 10 years from 1991
to 2000.[4]
In any case, the York Report compares this 5% ECGD return to the
return required more generally on Government trading funds of
6%, to estimate an implicit 1% subsidy. Crucially, it is our contention
that the rate of return should be risk adjusted. We would contend
that equivalent risk adjusted returns on equivalent business written
in the private sector banking markets would be substantially higher
than 6% real.
3.4 The 2000 NERA Report commissioned by
ECGD concluded that the ECGD's ability to offer competitive premium
rates could be an indication of subsidy, suggesting that there
needed to be a study to identify the costs to the Government of
using its balance sheet to support ECGD activities.[5]
The report strongly suggests there was a subsidy (see appendix).
In response, the Treasury and ECGD commissioned NERA to report
further on such a subsidy to ECGD. This NERA report uses the VAR
model to estimate the subsidy to ECGD and was published in early
2003.[6]
It did not justify VAR as against alternative approaches, and
came up with a series of results for the subsidy based upon differing
scenarios. None involved a significant annual subsidy of more
than £50 million. It relies heavily upon notional "K-values",
the ratio of unexpected to expected losses on deals, to cope with
risk. It accepts the weakness of the approach:
"the robustness of the k-values is subject
to a significant degree of uncertainty, which partly reflects
the difficulties of estimating k-values at high confidence interval
levels for "lumpy" exposures." (page 27)
3.5 The report attempts to overcome the
uncertainties of estimating K-values by claiming that such uncertainties
would suggest higher values within their confidence parameters
(between-100% and +100%). But they have failed to justify the
confidence parameters, and indeed there is no reason to suggest
that K-values could not exceed 100% (ie unexpected losses could
be significantly greater than expected losses).
3.6 None of the reports above reconcile
their conclusions to market risk premia charged in the international
bond markets. If the UK Government is so good at managing sovereign
risk then perhaps it should consider speculation in the international
bond markets: on the basis of the NERA report they would make
a killing!
4. OUR ALTERNATIVE
APPROACH
4.1 It is puzzling to us is that the York
Report and NERA chose to use VaR at the organisational level to
estimate the ECGD economic subsidy when they were aware of a much
simpler and reliable alternative, outlined in an Oxford Research
Group report released earlier in 2001.[7]
Reference to our paper is made in the full study, and is dismissed
as relying on "inferences from very different market activities".
We should like to explore this challenge below.
4.2 The ECGD provides finance facilities,
credit insurance and overseas investment insurance to UK based
companies. The bulk of these services entail providing guarantees
to banks and/or corporates against non-payment by an overseas
buyer (usually a foreign government) on long-term loans extended
to that buyer so as to facilitate British exports. Whilst the
detail of this relationship varies, the ultimate risks being insured
relate to interest rates, foreign exchange rates and sovereign
default.
4.3 Other risks, such as contract risk (ie
the possibility of a dispute over the quality of goods) are also
covered (under the EXIP scheme), and increase the exposure of
the ECGD. We have no satisfactory method to quantify the value
of these contract risks, and therefore valued them via an arbitrary
1% additional required return, or £46 million per annum.
4.4 The value of the sovereign guarantees
to the supplier or commercial lender may be readily ascertained
from the international debt markets. Foreign governments borrowing
dollars, for instance, pay a risk premium over the equivalent
American Government bond rate. This reflects the fee levied by
a commercial lender for taking on the risk of lending to that
foreign government. The value of ECGD guarantees is therefore
the difference between what the commercial lender would charge
the foreign government had the loan not been guaranteed, and the
equivalent rate were the loan to be made to the British Government
(ie substantially free of foreign sovereign risk). We use the
American Government bond rate as a proxy for the British Government
dollar borrowing rate, since the dollar is the most liquid sovereign
debt market, and a British Government dollar loan rate was not
immediately available (although it could be calculated). This
risk premium applies to the total amount guaranteed by the ECGD
in respect of each foreign government over each year the loan
is outstanding.
4.5 The total amount at risk at 31 March
2003 was £20,882 million, of which 22%, or £4,594 million,
was defence-related. We estimated an aggregate market risk premium
of 3.5% by noting that typical medium term emerging market bond
risk premia were at that time between 1% for South Africa to 6%
for Brazil. It should be noted that ECGD credit exposure is on
long term contracts to areas of high political risk, so the actual
figure would be well above the middle of the range. With more
information available in the latest ECGD Annual Report it is now
possible to make a better estimate of the market risk premium
for each guarantee, though to do this accurately would require
information on the repayment profile of each guaranteed loan—information
not presently made available publicly by ECGD.
4.6 From this gross cost of £138 million
we deducted the premium income received from companies of £34.6
million[8],
added £46 million for contract risk (see above), and added
an additional interest rate subsidy (averaged over last 10 years
at £3 million), to arrive at a figure of £149.4 million
for ECGD subsidies. In the light of ongoing research we are convinced
this estimate is conservative.
4.7 To the extent that the ECGD covers interest
rate and foreign exchange rate risk, equivalent instruments are
readily available in the derivatives market. The ECGD itself now
lays off much of its risk in these areas via established derivatives
markets.
4.8 There is also an additional market that
is fast developing that challenges the remaining function of the
ECGD. The credit derivatives market is a market for corporate
and sovereign risk, and is growing rapidly.[9]
It is a highly flexible market, and though young, has many options
for customers seeking to hedge their risks.[10]
The pricing of credit derivatives is based primarily on bond yield
spreads—the same method as above, though like other derivatives,
credit derivatives are very flexible financial contracts in that
their payouts can be derived from loan or bond values, default
or credit events, credit spreads, or credit ratings.[11]
These reference assets, in turn, can be associated with single
names, baskets or indices with cash settlement or physical delivery
of a relevant underlying asset or portfolio of assets. Indeed
in a recent article the use of credit derivatives to cover sovereign
risk was explicitly addressed.[12]
Has the time come for the ECGD to re-insure its sovereign risk
exposure via these markets? To do so would make explicit for all
to see the true quantum of the subsidy it gives.
4.9 It has been said that ECGD is involved
in supporting markets that the private financial markets are unwilling
to cover. This is true, though for a minority of ECGD guarantees.
The reason private financial markets are undeveloped in these
areas is because the risk is judged too high to give a price that
would both be acceptable to the exporter or recipient government
on the one hand, and adequately cover the risk on the other. This
does not in itself indicate a market failure, so much as a market
where demand at the market price is close to zero. In such circumstances,
the government may be justified in intervening to create a market
if it judges there are political or social reasons for doing so;
but it must be recognised that these are not economic reasons.
5. WHAT IS
THE COST
OF THESE
GUARANTEES TO
THE BRITISH
GOVERNMENT?
5.1 Can the British Government beat the
international bond markets? If the Government has access to privileged
information enabling it to reduce the risk faced on such transactions
below that faced by the market, then such information would also
enable it to profitably exploit such a position by speculative
trading in the international bond market (or indeed in credit
derivatives). We consider this possibility to be conceptually
problematic, runs counter to the prevailing economic policies
pursued by this and previous governments, and would have far-reaching
consequences for much of public sector finances (not to mention
economics theory). Indeed, current government policy (such as
Private Finance Initiatives, Public-Private Partnerships and the
like)[13]
is moving in the opposite direction, apparently based on the belief
that the private sector is the best vehicle to finance many areas
of the domestic public sector such as health, housing and transport,
where government control is so much greater. If it is appropriate
to involve the private sector in these operations, why is it appropriate
for the public sector to be engaged in guaranteeing private sector
(arms trading) operations?
5.2 Would the removal of such implicit subsidy
benefit the Government? As the ultimate guarantor of such debts,
the Government pays for such subsidies by an incremental deterioration
in its debt rating with the markets. This cost is currently "hidden"
across the full spectrum of Government borrowing, but the principle
is well established.[14]
The total cost to the exchequer of such subsidies can only be
reliably estimated using a market benchmark approach such as the
one we propose.
5.3 Neither subsidy calculation accounts
for severe or catastrophic risk, presently covered by the policies
issued by ECGD and ultimately born by the Treasury. The possibility
of extensive sovereign default by recipient states is real because
the ability to pay for imports is closely linked to the health
of the regional and global economy. That is, the risks are not
independent. The 1999 economic crisis in south east Asia ripped
through the region causing stock market crashes and extensive
defaults. Debt repayments were rescheduled, but the default consequences
have yet to run through the financial system. ECGD has not yet
written off many of the debts that are inevitably uncollectable.
A deeper crisis at a global level would present the sort catastrophic
event that would lead to massive losses by ECGD and the Treasury.
5.4 K-values certainly do not account for
catastrophic risk, but then neither does our market-based model.
Facing a catastrophic event, a private insurance company always
has the option to declare bankruptcy. Existing subsidy calculations
will therefore underestimate subsidies arising from existing policies,
unless those policies are altered to exclude such catastrophic
events.
5.5 Such an exclusion is unlikely, as the
whole banking system depends upon government guarantees against
bankruptcy. The banking sector anticipates that the government
is likely to bail out individual UK banks rather than see them
go under, allowing them to accept higher risks than they would
otherwise contemplate. This phenomenon, termed "moral hazard",
is well known. Because this moral hazard involves support for
exports, it entails a subsidy of risk that ultimately benefits
exporters and their customers. It is difficult to even guess at
the impact, and we have not attempted to include it, further implying
that our method underestimates the subsidies involved.
6. WHO BENEFITS
AND LOSES
FROM THESE
SUBSIDIES?
6.1 We believe it to be economically axiomatic
that subsidies can be justified when they effectively negate market
failure (ie where there are externalities,[15]
or where the market is inefficient[16]),
or for social or political reasons.[17]
6.2 Between a third and a half of ECGD guarantees
are made to the defence industry.[18]
Most ECGD guarantees are made directly or indirectly to larger
exporters in the business, in order to minimise administrative
costs. Guarantees are generally granted to a small number of the
largest companies that are well connected and established. It
would appear unlikely that the beneficiaries of ECGD guarantees
are organisations worthy of government business support on economic
grounds. Of course smaller subcontractors do benefit indirectly
from larger deals, but very much at the whim and control of the
prime contractor, who is able to determine the terms and conditions,
and achieve much of the profit.
6.3 Market distortions harm other businesses
and alternative industries that do not receive subsidies, through
greater competition for investment, skilled labour and other inputs,
and from worse terms of trade. The overall cost to British industry
in terms of reduced productivity and competitiveness from this
distortion could be considerable, but are conceptually difficult
to quantify. The market distortion implied by the Government using
its balance sheet to guarantee overseas sovereign debt is also
likely to come at an economic efficiency cost greater than simply
the incremental borrowing rate.
6.4 Who benefits from this system other
than the UK arms supplier? The overseas recipient and the UK exporter
clearly receive an economic benefit. The largest winner may actually
be the UK Bank as they clearly receive a loan markup at negligible
risk.
6.5 What is the impact of ECGD activities
on the recipient? The effects on the social and environmental
fabric of large construction projects or arms supply is well documented
elsewhere. The economic impacts of foreign direct investment crowding
out domestic investment or industry development can also be significant.
It is highly questionable whether the sorts of activity supported
by ECGD helps or hinders the government's objectives pursued by
DfID in trying to deliver sustainable economic development based
upon the nurturing of domestic industries. ECGD now takes into
account issues of sustainable development in considering projects,
at least partially, unless they involve defence contracts. Defence
contracts only scrutiny is through the formal DTI arms export
licensing system.
7. CONCLUSION
7.1 ECGD is pricing its products at well
below the market level (as measured by market sovereign risk premia).
This may be because of competing subsidies by other ECA. However,
the existence of a competitive market between governments in the
provision of subsidies is not an argument for subsidy itself.
Such a subsidy can only be justified on explicit political or
social grounds.
7.2 If there were no subsidy, there would
be no reason for the Government to maintain ownership of the organisation,
and proceedings ought to be commenced to privatise operations.
Indeed if the prime reason for the continued existence of the
ECGD is economic, thought could be given to providing explicit
subsidies to existing international banks to continue the role.
This would at least have the benefit of transparency.
7.3 Choosing where to allocate scarce Government
resources is a political decision. However, the allocation of
resources is frequently based upon prejudice and lobbying by interest
groups rather than informed advice to ministers. As numerous government
statements in the House concerning the financing of the Defence
Exports Services Organisation within the MoD demonstrate, Government
policy to support the defence industry has in the past been based,
at least in part, upon the assumption that this support brings
financial benefit to the Exchequer, largely through reduced prices
for MoD procurement from British companies. Support by ECGD for
exports, defence and civil, has been based on the assumption that
the cost to the taxpayer is minimal, while the benefit to the
defence companies concerned is significant. Why else would this
business have been retained within Government when the short end
of the business was privatised in the early 1990s?
1 ECGD, Annual Review and Resource Accounts 2002-03.
pp 7&8. Back
2
Estimating the Costs and Benefits of ECGD, National Economic
Research Associates, 2003, available on the ECGD website (www.ecgd.gov.uk) Back
3
Of course estimates of volatility might be estimated from the
international bond markets. However, to do so would simply be
to use our method indirectly. Since Chalmers et al consider that
benchmarking against such markets is of itself wrong, it would
be inconsistent for them to then derive their volatility estimates
from such markets. Back
4
The Subsidy Trap, p 21. Back
5
Estrin, S, Powell, S, Bagci, P, Thornton, S, Goate, P, The
Economic Rationale for the Public Provision of Export Credit Insurance
by ECGD, (2000, National Economic Research Associates, London)
Cmd Paper 4791; available on the ECGD website at: http://www.ecgd.gov.uk/nera.pdf Back
6
Bagci, P, Powell, S, Grayburn, J, Kvekvetsia, V, and Venebles,
A, Estimating the economic costs and benefits of ECGD; a Report
for the Export Credit Guarantees Department, (2003, NERA,
London); available on the ECGD website at: http://www.ecgd.gov.uk/neraiifinalreportjan2003.pdf Back
7
Ingram, P and Davis, I, The Subsidy Trap: British Government
Financial Support for Arms Exports and the Defence Industry, (2001,
Oxford Research Group/Saferworld, Oxford). Back
8
Annual Report, p 24. Back
9
Morgan Stanley estimate the CD market to be just under $1 trillion
as at mid-2002, "Spotlight, Credit Derivatives" in The
Treasurer, May 2002, p 45. The 1999 KPMG report for ECGD (para
4.7.6) and The Treasury itself identified credit derivatives market
as a potential method of ECGD hedging its risk in the market. Back
10
Frost, J, "Corporate Uses for Credit Derivatives",
in International Treasurer, 31 March 1997, http://www.intltreasurer.com/corpcder.htm.
"Like other derivatives, credit derivatives are very flexible
financial contracts in that their payouts can be derived from
loan or bond values, default or credit events, credit spreads,
or credit ratings. These reference assets, in turn, can be associated
with single names, baskets or indices with cash settlement or
physical delivery of a relevant underlying asset or portfolio
of assets." See also Davies, J, Hewer, J, Rivett,
P, The Financial Jungle-A User Guide to Credit Derivatives,
(2001, PriceWaterhouseCoopers, London). Back
11
"Spotlight, Credit Derivatives" in The Treasurer,
May 2002 p 49. Back
12
"Spotlight, Credit Derivatives" in The Treasurer,
May 2002, p 53. Back
13
It may or may not be accidental that the one common theme that
the ECGD has with such schemes is that they enable the Government
to receive monies now (whether capital sums or insurance premia)
in return for a promise to return significant resources to the
private sector at a later date. Back
14
In a similar vein, central Government restricts the capacity
of local authorities and other debt issuing apparently independent
organs of government to raise debt. And finance lecturers will
be well versed in responding to the old chestnut "why not
use the cost of debt to evaluate a project if the corporate chooses
to use debt to finance a new project". Back
15
Many externalities are of course social or political in nature. Back
16
For example, due to inequalities in market power, market information
or market depth. Back
17
For example, as part of a regional regeneration effort. Back
18
See ECGD recent Annual Reports, where a breakdown is given on
the proportion of business that is defence-related. Back
| 
