Internal Model Approach for New Credit Risk Regulation

Internal modeling in association with a new Capital Accord (part I)
Will segmentalize banks to a handful of international giants and the other thrift institutions

Considerable impact in comparison with the current market risk regulation

To summarize the Basle Committee's proposals for a new Accord (Table 1), its aim is to differentiate the capital requirement of an individual financial institution based on its borrowers' credit ratings. In this context, a new Accord will be called the "third Basle Accord," or "credit risk regulation.";

The third Basle Accord (a new Accord) will give a greater impact on the operation of financial institution than the previous one (the market risk regulation in 1996). Contrast to the previous regulation's mere effect in below-the-decimal percentage on the net worth ratio, a new Accord can force the financial institutions to increase equity capital by the unit of percent.

In the background of the Basle Committee's decision of review lies a gap between the existing Accord and the reality. The current Accord applies a flat credit conversion factor in the weighting of the risk asset based on the banking book (e.g., sovereign, bank, corporate) regardless of the borrower's rating. This came under criticism like "Isn't it unreasonable that the capital requirement for claims on the Hokkaido Takushoku Bank accounts for only one fifth for those on the General Electric, just because it is an OECD-incorporated bank?" Also, the growing number of off-balance sheet transactions has profoundly changed the world's financial markets. The advocate of closing the regulatory loophole by the use of holding companies in the Europe and U.S. must be worth listening. Basle Accord was originally created as preventative measures against the systemic risk on the global scale after the failure of the Bankhaus Herstatt and Franklin in 1974. Thus, the internationally active banks shall observe its rules.

Regulations might have caused the economic turbulence?

In the reviewing process for the new Accord so far, however, we have seen a jarring note thrust among the involved nations. The issue of the consultative paper itself delayed because of the counter-argument by the German authority. In Japan a criticism lingers pointing to the dogmatic assertion of a particular nation in the framework. On the other hand, more than a few voices prevail in the U.S. that the review takes too much time.

One reason for the criticisms is that the growing number of people are starting to doubt if the current Accord on a bank's capital ratio may have affected in a negative way on the recent economic dislocations in Asian countries including Japan. The concept of "provisioning the capital for the risk" is quite natural.

However, you must remember the technical term of "fallacy of composition" in the economics glossary. Well-capitalization of individual financial institution for a rainy day would not necessarily contribute to the stability of the world financial system. For instance, see what's been happening for the past several years since the prompt corrective action was taken. Every bank scrambled to pull back on lending even at home or abroad. We do not deny a merit of the Accord on the possible shake-out of institutions in poor operational efficiency. But then, still the doubt remains that the capital requirement may have triggered large-scale credit contraction and deflation - the tightening of the credit crunch. If this hypothesis is turned out to be true, we have overlooked the credit-multiplying effect of the Accord, i.e., the negative impact on the money supply.

Our suggestion is that the capital requirement should be more flexibly settable along with a current economic situation, in the same manner as monetary policies are subject to change according to the business cycle. Suppose if the U.S. economy is going through a serious deflation or a deep depression hit in the future. Is the world continuously capable of being in line with a new Accord.

Backlash against the only one winner

The second reason is the skepticism that behind the intention of tighter capital requirements the interest of any particular nation might be unusually favored. Recent series of Asian crisis revealed that the abrupt flow of the excessive short-term fund is quite devastating to the regional economy rather than contributing to its stabilization. Isn't this the "systemic risk" to be avoided according to the 1988 Accord? In addition, the internationally active banks are not the only occurrence factor of the systemic risk. The hedge funds must be counted in as a major threat.

However, the U.S. authority is reluctant to pose greater regulatory controls over the hedge funds at home. This attitude is casting a doubt that "the U.S. is giving priority in protecting domestic firms rather than in preventing the possible systemic risk in other countries." If the intent of a new Accord is anti-systemic risk, why the hedge fund is excluded?

The third reason is the review of the result of the second Basle Accord (market risk regulation). The market risk regulation is highly regarded by many regulatory authorities. On the other hand, it is widely believed in the financial community that administrative discretion has resulted with much ado. A sample is the U.S. authority's bailout of the LTCM in 1998 without letting the case to the market force.

Even in the U.S., not a few financial institutions have taken the similar investment operation as the LTCM did. Did the U.S. authority impose an adequate penalty, for instance, by increasing the multiplication factor (a modified coefficient for calculating capital requirement by administrative discretion introduced in the second Accord)? If not, wasn't the bailout a sort of protectionism?

In this way, the proposals for a new Accord are being countered by increasing number of negative responses with frustration against the "America the only winner." The public comments due is March 31, 2000. We strongly hope the financial community will submit prudential comments to the Committee.

Greater impact on countries with predominating indirect financing system

The outlook of the final proposals is still unforeseeable, however, let's run a what-if scenario to project the impact of the new Accord assuming it is finalized as is. So many countries, so many roles of the financial institution and the effect would be different by each country.

The asset size of a U.S. institution is relatively small. Conversely, the indirect financing percent is low in the U.S. on the back of the advanced securitization and direct financing system. Also, various instruments contribute to the easier reducing of loan assets; government-guaranteed MBS (mortgage-backed securities) market for the housing loans, ABS (asset-backed securities) markets for the credit card and commercial real estate loans, and sophisticated derivative market. Of course the U.S. has layers of secondary financial institutions (e.g., commercial bank, thrift, S&L, MSB, credit union) supporting the indirect financing system, however, they are not covered by the Basle Committee's regulations (domestic banks).

On the other hand, in Japan and Europe with less developed direct financing markets, the indirect financing has been furthering their economy. A bank has a larger asset with lower ROA in general. Another characteristic is the high ratio of the secured loans. Unlike in the U.S. where the unsecured loan is usual, the unsecured loan accounts for just a few percentage in Japan. Looking at the current condition, the Japanese economy is still in recession where the corporates are generally assigned lower credit ratings compared to those in the U.S. enjoying the strong economy.

Under all these circumstances, what will happen once the new Accord is enforced? First, the shift in lending would occur out of the smaller corporates with low ratings into the high-rated major corporates or sovereigns and investments in government bonds. The proposals require a 150% risk weighting to the loans to a corporate rated below B-. This is tough enough. Japanese financial institutions that have long been in poor health would spur pulling back on lending to the small businesses.

Another significant problem is how to handle collateral. The proposals do not recognize the risk mitigation effect by real-estate collateral. It seems to go so far as to decide that security by the mortgage totally makes no sense. Setting aside its reasonableness, this recognition gives mere affection to the U.S. where unsecured loans are the primary force. Contrarily, this would work in significantly disadvantageous way to Japanese institutions which mainly rely on the collateral approach in lending. This would deepen the international competitive distortions across the financial institutions.

Here is the summary of the concerns over the Japanese financial institutions about the new Accord, if the proposals are adopted as they are; (1) further deterioration in Japanese banks' ratings and (2) further credit crunch and negative effect on the macro economy.

Survive as an investment bank or savings bank?

The proposals might be withdrawn. Yet, counter-arguments against them could work in just buying some time as far as the globalization continues.
It is clear for the Japanese institutions which way should they take - Americanization. This approach consists of following two options:

Option 1 - Money center bank or investment bank. Thrust into investment banking targeting the major corporates or markets as an international bank complying with the Basle Accord. Put the consumer and small business sections in major cutback. Or even downsize their operation unless the fee-for-service business pays out. On the other hand, get psyched and expand the overseas operation aggressively. You may make a competitive bank with high ROE.

Option 2 - Regional bank like the former Nations or the Bank One, or savings bank in Japanese style. Opening overseas branches is out of the question. Just comply with the domestic regulation which is more relaxed than the Basle Accord. With a low loan-deposit ratio, the equity capital is likely to become higher, which should be used to new lending. Invest the excess funds in various markets. Granted the high rating. You got a nice neighborhood bank positioning itself within intra-regional flows characterized by industrial development.

In the view of the propositions of a new Accord, it seems quite abnormal for the Japanese city banks taking a similar all-rounder strategy to go for the retail, wholesale, and even the overseas operation. A financial institution with lowering rating would be forced to follow the Option 2 as a domestic bank, even though it is big enough. If a bank could be reborn after acquisition, merge, or divestiture to make a major international bank as described in the Option 1 (maybe less than five banks in Japan), it would be capable of inventing its future having its influence revitalized.

We must emphasize the direction in specializing as a domestic bank (Option 2) is worthy too. The Japanese media's exaggeration of "the U.S. banking industry = money center banks = high ROE management making good money with highly advanced technology" is just an image created through the worship of America. You should visit any town in the U.S. and watch the banks on a street. The reality in the U.S. banking represents a layered structure. One layer consists of several major banks, each of which maintains a network of a few thousands of branches with employees several times larger than those of the Tokyo Mitsubishi Bank. And the other layer is the rich secondary financial system consisting of more than 9,000 institutions. Thus, investment bank makes the minority different from what the Japanese thinks of.

Back in Japan, it seems quite unusual that every financial institution is striving for the same managerial strategy of "Become an investment bank" while restructuring the already tiny network of less than 500 branches and cutting the personnel that has been in short.

From the perspective of sound banking and commitment in industrial promotion expected from financial businesses, international bank must not be the only one to be welcomed. To the corporates, international bank will rather be somewhat too upstage to have relationship and just to ask for lending to their businesses. It's high time to make an about-face in preference for brand-names like an investment bank or money center bank. Extremely, give a speech in a morning meeting declaring that "We're going back-to-basics in financial business. Now, everybody get ready for deposit-taking!" The personnel appraisal stresses on how much sincere services the employee provides the customer with. At the same time, increase staffing while holding down the wages to maintain low overhead cost. This must be wise and drastic measures. Recently we see many Japanese banks downsizing. However, the number of branches and personnel at the Deutsche Bank or the Citibank is several times greater than at the Tokyo Mitsubishi Bank. There can be managers of financial institutions who recognize "the quality service do rely on human power after all." To ensure the competitiveness, they should retain the IT strategy in selecting the borrowers and maintaining the standard of services.

Reviewing of the financial system becoming more critical

The Japanese financial industry should bring to the government attention the fact that the Japan is substantially behind the U.S. in the financial system, and ask for urgent improvements to avoid possible disbenefit under a new Accord. You can't waste years to discuss a minor issue like repeal of the securities transaction stamp tax when the enforcement of a new Accord is coming only in two years at the minimum.

For instance, it is a serious disadvantage there is no MBS and ABS markets in Japan for shedding assets by converting them into securities. Why in Japan the housing-loan securitization has not come off well contrasting to the penetration of MBS in the U.S.? One reason is the existence of government agencies (e.g., Ginnie Mae, Fannie Mae) that give the government guarantee on the housing loan claims of a bank. The widespread floating rate housing loan in Japan can be converted to an adjustable rate mortgage (ARM), if it is backed by the government. Holding such ARMs in a fund would improve the yield, and provide a good investment on the personal assets in poor performance. You should start lobbying for these issues right now.

Once getting ample supply of matured on-balance sheet products with government guarantee, you can produce a variety of financial instruments held with off-balance sheet products. You may also hit a bunch of ideas in repackage products in combination with the deteriorated commercial real estates (such as the securitization products on Latin Americas debts in which the U.S. banks do well).

The U.S. takes advantage in their time-honored on-balance sheet item which has been a source for good financial instruments - strips bond, at top rating with a long-term discount. After the two long-term credit banks have gone in Japan, regulations in place restricting bonds with lifetime of five years make no sense. Once you get the interest-bearing government bonds, decompose it in the form of individual certificates one after another for circulation. This could resolve the long-lasting headache of liquidity shortage in the secondary bonds market.

For long time the Japanese banks' product development sections, suffering from the regulative uncertainty and opaque administrative guidance, have stood the media coverage blaming that "the Japanese banks have no capability of developing a new financial instrument because they're so far behind the U.S. banks in financial technology." Now is the time to wipe off this stigma.

Also, the balanced application of capital ratios to the domestic banks is expected. The lower the capital ratio, the better it serves in the economic growth by reducing the cost of the equity capital, as far as the risk is under control. Things are completely different in the international settlement such as by the World Bank or the IMF with a limited ability of coordination and in the domestic market where a central bank functions as a buffer. If certain endeavors are taken in preparing an appropriate safeguard against external factors like hedge funds to enhance the whole market's ability to withstand risk, the capital adequacy ratio for the domestic banks must be more lowered. The reasons for a bank's public responsibility or role in the development of local economy can be favored over the equity rationing by rating as far as such reason will not interfere with the structural adjustment. This must be the raison d'etre of the financial system.

In the U.S., a new Accord is likely to close one loophole of circumvention via holding company. However, the other loophole of using derivatives would be still available. Some people say that about 25 percent of the U.S. banks' total assets has already get benefited in this way. You should point to the necessity for improvement in terms of equity concerns.

Supervision over the foreign banks in Japan is highly important. If the Japanese banks are continuously suffering in weak health, the foreign banks will gradually carve out the domestic market share while increasing their influence in the Japanese financial system. Letting them practice their own way of doing business as if they've got extraterritoriality pleading that "You can't rely on the Japanese voodoo economy" would be a significant threat to the stability of the Japanese financial system. Stop the implicit and opaque transaction of profit manipulation with the off-balance sheet item via the foreign bank. We only hope the equal opportunity will be set up in line with the domestic system, practices, and policies while the deregulation and open-door policy are maintained.

Using internal model becomes essential for a international bank

Concerning the use of internal models in the credit risk quantification, the Consultative Paper "commends the use and continued development of internal models," however, remains in cautious tone of "still need to be cleared before these objectives can be met." Among the international financial community, particularly the U.S. industry groups are keen to adopt the internal modeling. President Macdonough at New York Federal Reserve Bank seems siding with them. We still have the time before enforcement of a new Accord. Recalling the time when the existing Accord took effect, you cannot eliminate the possibility in adopting some internal modeling technique.

Not to mention the U.S., and even in Japan the hearing process started as a part of monitoring on international banks management concerning the "usage of internal model in the credit risk calculation," for instance. An international bank with no internal model should get prepared to obtain severe assessment from the future inspection by the Federal Reserve Bank. In other words, to the internationally active banks under the influence of the U.S. Federal Reserve Bank or the England Bank, the usage of the internal model is becoming mandatory regardless of its JOUBUNKA in a new Accord by the Basle Committee. The race is already started.

The Part II will present a detailed analysis concerning the internal modeling issues.

Internal modeling in association with a new Capital Accord (part II)
Utility and limit of portfolio-based credit model

CreditMetrics, the only readily available model at present

First, you need to know only three models (Table 2) which are applied to the "portfolio-based credit risk model" in a new Accord. Among them, you should remember CreditMetrics [by Risk Metrics Group (RMG), a risk model provider separated and independent from J.P. Morgan] as the only available model at this time that has the true utility in the practices. You might feel confused of several other models also called "credit risk quantification model" (e.g. default rate estimation model, credit derivative valuation model, potential exposure calculation model, corporate bond valuation model). Note that these are totally different ones from the internal model defined in the proposals.

Of course there are lots of models other than above including the ones developed by the Japanese banks. Notably, some of them are more excellent than the models listed in Table 2. However, as described later, the step of measuring the contents with each other seems to have taken the backseat to the argument about which one should be used as the standard. Any model not yet visible on the international horizon is ignored at this stage whatever merits it may have.

Actually, these three models are neither pioneer nor original. For instance, the immediate ancestor of CreditMetrics is Portfolio Manager (released in 1993) by KMV. Basically, Portfolio Manager was developed as a risk quantification model of listed companies, which was an acrobat for application to a bank's portfolio management including claims on the unlisted companies. To supplement for this need, many improved models were presented by the banks all over the world. CreditMetrics is the most famous one (released in April, 1997) among such modifications.

Additionally, both Portfolio Manager and CreditMetrics have their basics rooted in the Merton's model (in 1974) that valuates the stocks as call options to corporate assets.

CreditRisk+ was released at a successful timing (in October, 1997) after the CreditMetrics brand had been penetrated. CreditRisk+ is appealing by its simple model using the Actuarial mathematics technique to overcome the CreditMetrics' weak point in huge computational complexity. CreditPortfolioView, launched at the same timing (autumn 1997) as CreditRisk+, is capable of quantifying the credit risk using the econometrics model.

For all those models, the result we estimate CreditMetrics' exclusiveness is based on the reason that the rest has clear problems. In case of CreditRisk+, its rationale of Actuarial mathematics can function effectively only under the limited conditions characterized by (1) massive loans and (2) no concentration on the part of the loans, such as in the credit card or housing loans. So, it can be easily proved that CreditRisk+ would make underestimation of future risk for a Japanese bank's portfolio if it contains huge non-performing loans to a general contractor. For details, see the documentation on our website at http://www.numtech.com/.

Though running out of steam, CreditRisk+ is still a comparison counterpart because not so many people have the experience with internal models in their line of business. CreditPortfolioView's econometric approach is so basic that it may be tied to what the proxy value of the corporate credit fundamentals should rely on. The CreditMetrics designers must have hit upon this idea too, however, they didn't incorporate the econometric approach in CreditMetrics due to its well-known major disadvantage (frankly, it won't hit). Reversely, the easiness of incorporating econometric elements in a CreditMetrics model will galvanize the coming out of a hybrid model.

Under these circumstances, the current argument on the internal model seems to have an inclination toward CreditMetrics approach.

Three truths over internal modeling

Now, let us clarify the three basic facts.

First, it is not possible to obtain the true measure of the credit risk. Even the market risk models couldn't expect the lasting turmoil on the markets after 1997, which brought as huge losses as several U.S. banks were forced to be sold out. Enumerating the economic crises in the past 10 years - Black Monday, sharp rise in U.S. interest rates, Latin American debt crisis, Scandinavian crisis, break up of the ERM, Asian crisis, Russian crisis, and recent Japanese financial disturbances - every theory, from option model to interest arbitrage model, had been overwhelmed in these waves of upheaval. Certainly, models have been constantly improved, however, there is no guarantee that the currently effective model would work out again in the future. Moreover, the credit risk is supposed to be handled in a unit of several years. If the economic structure changes during this interval, the model would be logically demolished. Therefore, the internal model in a new Accord must be best-effort base.

The importance of backtesting is also stressed. The true function of backtesting both in the market risk model (the second Basle Accord) and in the credit VaR (a new Accord) is to provide requirements for eliminating arbitrariness and retaining objectivity. So, it would be a critical misconception that the backtesting is usable as a model validation tool. The fact that nobody can know the true credit risk amount has put the argument on the regulation into a sort of proofless trial.

Second, the credit exposure contains regional factor to a large extent. This makes the point that is apt to be overlooked when discussing the credit risk in line with the market risk. Today most of the marketable instruments are internationally transacted and the trading system is converging to the one that had been tried in the U.S. It is considered that the market risk exposure of an internationally active bank can be measured and compared with each other by a single model. However, the credit risk exposure will not go the same way. Under different jurisdictions, the default has the different meaning and the post-default behavior of the model should vary. However high market shares S&P and Moody's are holding, the rating system is applicable only to the major corporates in the major industrial countries. Focusing on the G10 countries, the economic structure extremely diversifies by each nation, which would not allow a single behavior of the credit risk even after taking correlation factor into account. The Committee's new Accord seems like a square peg trying to fit into a round hole in such diversification, intending to extend its application even to non-G10 countries. The idea must be too grand, though its significance is recognized in a way.

Third, not all countries are set on common ground with a new Accord. In the U.S., thanks to its enhanced market transparency, good amount of data (e.g., rating, spread, correlation) are available in high reliability which can be incorporated in the model. On the other hand, the efficiency in credit markets in non-U.S. countries is relatively inferior. This opacity depends on the lack of advanced accounting system, which causes window dressing. Further, it strongly reflects the social structure with factors of uncertainty such as regionality, industrial affiliations (keiretsu), history, clan and lineage, religion, ethnic concerns. Probably unimaginable from the younger nations' view, the reality in the time-honored system is, for instance even in Japan, that you have to worry about the possibility of racketeering in the process of debt collection. Fitting a mathematical model in a country with less transparent social system will face the difficulties. Thus, there's no saying that the criticism of "slow response to internal model approach" is whether to blame the target nation for its incapability in preparing the model or to disapprove its social structure different from the criticizer's nation.

Supervisory approach beyond authority's capability?

The Committee might just as well pay open-eared attentiveness to a conservative stance responding to the above problems. What makes them to press the internal model approach? It is based on following three reasons:
First, if the authority is persuaded to dichotomize the banking sector to internal model based banks and non-internal model based banks, the former would be able to continue their businesses on more advantageous rule. In fact, recent comments by the FRB sources show they are more likely to be getting convinced of this direction.

Second, the use of internal model is seen as a strategy of getting higher rating of the financial institution. The development of a de facto standard would add further credit. The better the rating of the institution becomes, the more favorable it can keep the edge in the international competition (e.g., the Bankers Trusts' RAROC case). It would also appeal to the rating agencies. Moreover, the institution is even making a profit by selling the internal model to outside users. The internal model has effectiveness like advertisements on the Formula One racing car.

Third, the credit risk quantification model would link to the pricing model of credit derivatives like heads and tails of a coin. The key barriers in selling the credit derivatives are the opaque pricing and the lack of liquidity. Therefore, once the credit risk can be measured, you'll see a huge credit derivatives market emerging on the horizon. This sounds quite alluring. As in a few decades back in the U.S. when the mortgage business was rising along with the shifting in money demand from the heavily industrial area of the Northeast containing steel and motor factories to the newly emerging industrial area in the West Coast under the slogan of "from Rust Belt to Sun Belt," everybody is now looking forward to the possibility of the credit derivative business utilizing the regional divergences in the credit exposure.

The Committee's internal model approach can be the preparation for this potential. Setting up of the regulatory arbitrage market, a financing approach to improve the capital ratio, is also attractive.

With all these reasons, the industry groups [e.g., International Swap Dealer Association (ISDA)] or the influential industrial magazines (e.g., Risk) are making a pari passu advance in the request of early adoption of the internal model approach. Anybody who now claims negative opinion in every hue of holdback would get labeled a "wimp."

However, the ways in the promotion of this issue to this stage seem unsound. Take the case when setting the second Basle Accord for example. Distinctly, it was necessary to introduce the internal model approach at that time. Then what was the result? The supervisory approach was so beyond the authorities' capabilities that they had to buy all expertise in model monitoring from outside accounting firms. This produced an out-of-sync situation that each institution develops the VaR model on its own discretion and the accounting firm with less expertise in this field checks over it. Surely the transparency may have been improved within an institution. However, only god knows how objectively the capital adequacy requirement is imposed over every one of institutions which diversifies both in nationality and nature. Uncritical acceptance of what the second Basle Accord brought in can be said too optimistic.

After all, strong concern remains that sophisticated internal model, if permitted without matured discussions at length, would be used just as a loophole by the internal model based institution. The same can be applied in the setup of a guideline for introducing the internal model. The aggressive drafting of the international standard without reaching a consensus would result in discretion of each nation, possibly inviting a hollowing out of a new Accord.

Misunderstanding of U.S. advantage and Japanese disadvantage in banking system

Regardless of the fact that many of the major Japanese banks have estimated the credit VaR using the internal model, they receive a wrong press covering such as "the Japanese banks are incapable of self-developing the internal model" even in the main financial daily still now. It is regrettable that some banks' efforts in heavy promotion on their websites or annual reports seem to have turned all in vain. However, it would be too namby-pamby for researchers on the theoretical model taking an attitude of "some day truth conquers all things."

They must understand the discussion on modeling should take a realistic approach such as standardization of video tape or cellular phone rather than quest for a scientific truth. Or the other expression of "those who command the standard will predominate in the market" would be a paraphrase. So, let us offer some suggestions to cope with this issue.

First, recognize the fact as exactly as possible. Internal modeling is quite challenging. It is considerably number-crunching far beyond the market VaR. A facilely-built system would make a model that takes several days in completing a simulation. The data maintenance is another labor to do. Making a universally-applicable model will be quite tricky. All these problems root in an unanswerable question of how this uncertain world is compressed into a model rather than shortage of labor. Keeping up with the fast business operation is also a great deal. Unlike marketable instruments, the credit exposures management is hardly standardized due to the strong color of regionality and individuality. Apprehensions remain about how a gargantuan bank evolved from international merge would build an exact model for their risk exposure.

For these reasons, even the U.S. banks are making little progress in internal control. The road to the utilization of internal model climbs steeply up partway where the sophisticated banks seem to have stagnated bumper-to-bumper. The practices of credit risk model in the U.S. banking sector are "disappointing" as commented by Governor Meyer of FRB (June 3, 1999), even at the money center banks. At the so-called sophisticated financial institution, contrast to the model development section's exploitation of "our own-developed model is the best," the people in the business line assess it "not nearly so good." There is no commercial system abroad that proves utility (just with much ballyhooed effort).

Second, you should recognize the internal model can serve in risk management and tackle it positively. As pointed out above, using an internal model is as effective as a billboard in the Formula One racing. Probably it can act as a far better imagemaker than forming a sport team having the bank name. It is also useful as a foundation technique for credit derivatives. After all, you can use the internal model as a weapon rather than just a firewall against the risk.

Citing the movie "Wall Street" in my evaluation, the phrase "greed is right" seems right at least in the Anglo-Saxondom. The Japanese banks should show an aggressive side, not always being draped themselves in the Confucianism. If the management honestly pursues for higher level ROE stated on the improvement plan, there is no instrument other than relying on aggressive derivatives. Furthermore, this could be an ultimate countermeasure against a new Accord aiming at its hollowing out.

In addition, if the internal model is as at elementary technical level as CreditMetrics is, it can become standard in compliance with the regulation. Then, incorporate the factor of local situation of each country in the model while correcting the weak point, the accuracy in the credit risk quantification can be improved (Table 3). Internal modeling is a topic in relation to the managerial decision in which way the business should be running. It matters a lot if reasonable assessment cannot be conducted at your bank on the rival bank's dubious loaning behavior or credit derivatives, if any. Also, quantifying your bank's data with the internal model will vividly visualize the reality of the deteriorating credit exposure in chart. In fact, the transparency in business decision can be secured at least.

Third, be more strategic. Unfortunately, insisting on conservative stance against the internal model approach may not turn about the trend at this stage. Then, make every effort to remain on the winner's circle at the endgame. The Japanese banks are no gamesmen and apt to take cosmetic approaches. Thus, their excellent human resources remain unutilized in a sort of embarrassment of riches. In fact, the Japanese banks have staffed the risk management section with best and brightest people, and some of them even had developed internal models long before the release of CreditMetrics. It is quite regrettable these fruits are not expanded into a weapon.

The internal model argument is no high-toned issue like a scientific research. The model approach supporter banks might be seen in somewhat a hit-and-miss fashion. You need a powerful propaganda and get armed with perfect theories. Form a team with aggressive staff as if their blood is made of Anglo-Saxon's (i.e., fire incompetent staff and headhunt external resources if necessary), then strategically form alliances with other banks irrespective of they are Japanese or foreigner. Command the checkbook journalism, pay moneys on the Ph. Ds. Play the game of your counterpart. You will not survive the competition among financial institutions on the global scale as far as you play the saint.

Conclusion

Through the two serials of this paper, the potential impact of a new Accord is pointed out as follows: (1) A new Accord will lead to the rational allocation of risk capital of individual financial institution, however, (2) will cause disintermediation of bank-system lending while leaving a possibility for a macro recession in nations where the indirect financing system is predominating. Also, this paper discourses on (3) the less maturity in the discussions on internal model for getting incorporated in the regulation. Taking all those into account, the way the Japanese financial institutions should take is suggested as follows: (4) reshuffling of banking sector to make a few international banks, (5) reviewing of the financial system to reform the market to the one that is capable of optimizing the assets or serving regulatory arbitrage as in the U.S., and (6) going the Anglo-Saxon way in human resources and managerial strategies to metamorphose into more combatant corporate groups.

Globalization of economy is an unavoidable trend. Under these circumstances, the business strategic uncertainty is reducing. In other words, a misjudgment will ask for a pressure to get out of the market at a very early stage. Managerial choices are decreasing. I only hope the Japanese banks will respond appropriately to a new Accord.