«by MICHAEL GUDGER FAO AGRICULTURAL SERVICES BULLETIN 129 The designations employed and the presentation of material in this publication do not imply ...»
There are no data available as to the cost of operations and the consequent charge to the budget. This information would provide a more complete picture of the programme and allow a balanced consideration of the sustainability without ongoing subsidies It is not possible from these statistics to know whether or how many of these loans would have been made in the absence of guarantees. However, the case for additionality at relatively moderate cost is certainly stronger in Canada than in any other country surveyed, possibly excepting Germany. Almost 40% of the borrowers are startups, an additional 10% are in the first year of operations, and about 63% are in the second or third years of operations. The average borrower provides about 4.8 jobs and according to SBLA, adds additional 2.4 jobs per loan. Thus, the vast majority of guarantees support start-up and young businesses employing fewer than 5 people, which can add an average of another 2.4 jobs per loan guarantee, issued. If additionality is found in any guarantee system, it will be found in Canada. Including administrative costs and the guarantee payments that should result from the enormous increase in the volume of lending, it will be possible to measure both the additionality and sustainability by applying cost benefit analysis for job creation and the subsidy dependence index to measure sustainability.
ASIAN GUARANTEE SYSTEMS
Credit guarantee programmes for SME and micro borrowers operate in many Asian countries.
However, the information available is even less comprehensive than in the case of European programmes. Only a single analytical article15 has yet to appear. This account draws heavily upon the data in that article, although the conclusions drawn are rather different.
Japan is estimated to have about 3.4 million SMEs. Loans outstanding in 1994 to these SMEs are estimated at US$3,560 billion of which only US$269 billion, about 7.5%, are guaranteed.
52 separate guarantee companies staffed by approximately 100 people each, issue these guarantees. These guarantee companies have a retro-guarantee from a government owned company for 70% of any losses that incur due to the total or partial failure of a borrower to repay the loan. This appears to be the principle conduit for continuing administrative subsidies. The guarantee industry reports 1994 income of about US$7.8 billion. However, only about 2.2 billion, 28% comes from guarantee fees. The rest is from investments and other income. It appears that the initial capital came from government and that much of the retained earnings are derived from the financial income on that capital. No financial data on the reguarantee account is available. Annually about 1.4% of the guarantees are called and of this percentage about 54% is recovered, producing a net loss of about 0.7% of the guarantees issued. The 1-% fees charged should therefore cover this cost and make a contribution to capital. The budget support for these programmes is US$68 million. This is about 2.5% of the amount of the loans guaranteed.
A calculation by Hatakeyama of the administrative cost in Japan per US$1,000 of loans guaranteed is US$28 or 2.8% and only marginally higher than the above calculation. The losses would appear to be offset by the fees charged. Thus, if the Hatakeyama calculation of the cost of operations were fully charged to buyers of guarantees, the cost would be about 3.5% (2.8% administration + 0.7% losses net of recoveries).
Michiko S. Hatakeyama and others, “Credit Guarantee Systems for Small and Medium Enterprises in Some Asian Countries” in The Financier, 1997. The Spanish version appears in Sistemas, 1996. The article contains what appears to be some confusion as to the total value of guarantees issued and the outstanding value of current guarantees.
Korea In Korea, US$72.6 billion of loans are made to SMEs but only about US$9.5 billion or 13% is guaranteed. Guarantees are called at the rate 6.8% and only about 23% of called guarantees are recovered, producing a net loss of about 5% of the amount of the guarantees and about 4% after the 1% guarantee fee charged is deducted.
Hatakeyama reports that it costs about $70 per $1,000, or 7%, of the amount of guarantees issued to meet operational costs. Despite these high costs, the scheme is expanding vigorously, apparently financing the difference between the cost of operations and guarantee payments and the fee income from earnings on their capital. As a consequence of rapid expansion, the guarantee payments have increased by nearly 400% between 1988 and 1992.
“If this trend continues, subsidizing the credit guarantee scheme may eventually become too costly for the government.”16 Government budget support is $3 million or about.032% of the US$9.5 billion of guarantees issued.
The combined administrative and loss costs of the Korean guarantee programme appear to be about 11% of the annual value of guarantees issued. While this is considerably higher than the cost in Japan, it is somewhat less than the costs found in Western European countries, excluding Germany.
Indonesia has a smaller programme. About US$15.8 billion of credit flows to SMEs but only about r $12.6 million, 0.08%, of this amount is guaranteed. A full 50% of the guarantees are called, of which 71% are recovered, for a loss of 14.5% and a net loss of almost 14% of the value of the guarantees after the 0.6% fee is collected. Indonesia supplies about $20 million per year from the government budget to support guarantees of only $12.6 million.
Malaysia has a government backed SME loan guarantee programme, the Credit Guarantee Corporation (CGC). The CGC was established in 1972 and services three sectors, small merchants, traders, and small construction. Bank Negara, the Central Bank, operates a quota system assigning loan amounts to each bank for each of these sectors. These loans have an artificially low interest rate ceiling. The various schemes that operate within CGC begin with a large number of loans and have declined in each subsequent year.
The General Guarantee Scheme (GGS) was introduced in 1980 to replace the defunct CGC.
It made over 6,000 guarantees that year; but in 1985 it issued less than 100 and was suspended in 1989 when it issued only 5 guarantees. The Special Loan Scheme (SLS) introduced in 1981 and made over 17,000 guarantees; by 1989, the annual volume had fallen to just over 1,000 guarantees. Finally, the Principal Guarantee Scheme (PGS) was introduce in 1989 and issued 5,641 guarantees in 1991 but only 352 in the first 2 months of 1994, when yet another version of the PGS was introduced.
These guarantees were generally issued without the knowledge of the borrowers. A survey found that most borrowers (26 out of 32) were unaware of their existence. The CGC apparently does not give a very high priority to paying its claims and this probably accounts for the declining volume of guarantees in all three programmes. From 1986 to 1993, the CGC processed 3,563 claims for Ringgits millions (RM) 36.4 but paid only 1,505 claims totaling only RM 9.3. The remaining was rejected on technical grounds or withdrawn by banks. In 1993, the latest year for which data is available, non-performing guaranteed loans were put at about RM 185 or about 34% of total loans guaranteed. However, the CGC had processed only RM 5.3 of claims and had paid only RM 2.4 of claims. This is actually a sharp improvement from earlier years - both in terms of defaulted loans and amounts paid to banks!
Grahame Boocock and Mohammed Noor Mohammed Shariff, “Loan Guarantee Schemes for SMEs—The Experience in Malaysia,” Small Enterprise Development, Vol. 7, No. 2, June 1996.
Philippines A recent count in the Philippines identified 10 different guarantee agencies, each with large administrative overheads and a very small volume of operations. Most of these programmes were so small that the guarantee income was not sufficient even to cover the administrative charges incurred by the Manila-based office staff.
The largest and most interesting Philippine guarantee institution is the Quedan Corporation (Quedancorp). It was set up to provide guarantees on rice warehouse receipts, thus making them negotiable. A miller, or farmer, can take his paddy to a Quedancorp supervised warehouse, deposit it, and use the receipt as collateral for a loan or take the receipt to the bank and discount it for cash. Quedancorp guarantees that the rice backing the receipt is warehoused and in an acceptable condition.
Between 1979 and 1993, Quedancorp income from guarantees was about 89.4 million Philippine Pesos (PHP) from guarantee issued on about 50,000 total transactions worth about PHP 15,166 million. The guarantee fee averaged across the period about 0.6% of the amount guaranteed. Over this period, the administrative costs of issuing the guarantees and supervising the warehouses was about 1.8% of the amount guaranteed. While guarantee fees were about PHP 90 million, the operating expenses were over PHP 270 million.
A large reserve created by the Philippine budget enabled Quedancorp to survive. Quedancorp investment income and loan interest exceeded 630 million Pesos, although inflation severely eroded the real value of its reserves. Beginning in 1990, the guarantee fees declined from about PHP 15 million per year to only about PHP 6 million per year in 1993 while the administrative charges rose from about PHP 35 million to well over PHP50 million. In the last year for which data is available it cost over PHP 50 million to issue guarantees that generated income just under PHP 6 million. In 1994, Quedancorp was reorganized and recapitalized, planning to discontinue the guarantee business to become an agricultural lender The pattern is much the same at the Guarantee Fund for Small and Medium Enterprise (GFSME). Between 1984 and 1993, the GFSME generated origination fees of PHP 17 million and incurred expenses of PHP 211 million issuing about 1,360 guarantees (about 136 per year) on PHP 2.1 billion. The investment income from the reserve funded by the Philippine budget offset the differences. Charges in excess of 10% of the amount guaranteed would have been required to cover GFSME’s operating costs.
The situation was much the same at the GFSMEs competitor, the Small Business Guarantee and Finance Corporation (SBGFC) which only began operations in 1992. Between 1992 and 1994 when the study was completed SBGFC had generated guarantee fees of under PHP 3 million and expenses of nearly PHP 26 million on 525 guarantee operations totaling PHP 470 million.18 One of the few other empirical studies of the Philippine experience with guarantee funds investigated several other guarantee programmes: the CALF guarantee programme to promote agricultural lending, the Industrial Guarantee and Loan fund (IGLF), and PhilGuarantee to guarantee coverage to Philippine exporters on their local and foreign borrowings. The PhilGuarantee suffered 95% calls on their guarantees and the company was wound up. The
study concluded that:
"The Philippine case shows that guarantee schemes of social banking are a type of support and are institutionally unviable. In actual fact they restrict, rather than encourage, lending and offer no solution to the problem of small and microenterprise finance. As they do more harm than good, only two options can be recommended: to leave risk management either to banks or to small and microentrepreneurs and their own grassroots solidarity arrangements." 19 Consulting report prepared by the author for the World Bank, 1994.
Hans Dieter Seibel, “Credit Guarantee Schemes in Small and Microenterprise Finance: Do they really do more good than harm? The case of the Philippines.” Quarterly Journal of International Agriculture, Vol.
32 No.2, April-June 1995.
SUMMARY OF THE ASIAN EXPERIENCEThere is still much research to be done on the Asian schemes, especially their financial arrangements and the total cost of the guarantees. However, it is reasonably clear that the pattern found in Europe is evident in Asia. The market for guarantees is relatively small.
Only a small portion of borrowers obtain guarantees. The administrative cost of issuing the guarantees is quite high. It is doubtful that borrowers would accept these costs if they were required to pay them to obtain guarantees.
An increase of 3.5% in Japan or 7% in Korea would be acceptable only to the most hardpressed borrowers. In Japan, the guarantee costs would exceed the interest rate currently charged on loans and in Korea would double the interest rate. In the Philippines, there is no doubt that most borrowers would not use guarantees if they were charged the full cost. The Malaysian practice seems little other than a thinly veiled asset transfer. Banks are assigned SME lending quotas at below-market rates on loans guaranteed by a government corporation which denies payment on 35-50% of the loans that default. In these circumstances the continuing state subsidy, as in Europe, is a key element. It allows the guarantee funds to market their product in low volumes at below cost, while simultaneously incurring very high administrative costs. The difference is funded either by a direct subsidy or by interest income on a capital reserve derived from the state budget.
Africa has dozens of guarantee schemes, most established relatively recently with donor funds and operating on a very small-scale. Meyer and Nagarajan_ provide an inventory of 20 schemes in 16 countries. However, the schemes are fairly new and data are either absent or so scanty that very few conclusions can be drawn. Meyer and Nagarajan conclude that several programs issued few guarantees and were terminated. Terminations occurred primarily because of poor performance and poor implementation, which led to high costs and defaults.
Further, they note that where guarantee schemes have been relatively active, they make little impact. There was little additionality in terms of loans made, and almost no borrowers "graduated" to non-guaranteed lending.
Agricultural Credit Guarantees of Nigeria
Nigeria offers the single African case where data are sufficient to permit some conclusions.