• 沒有找到結果。

RMBI Newsletter = 風險薈訊, Issue 2 (February 2011)

N/A
N/A
Protected

Academic year: 2021

Share "RMBI Newsletter = 風險薈訊, Issue 2 (February 2011)"

Copied!
24
0
0

加載中.... (立即查看全文)

全文

(1)

航空業

風險管理及商業智能工具的運用

RISK CONTROL WITH BUSINESS INTELLIGENCE TOOLS

IN AIRLINE INDUSTRY

信貸風險

Credit Risk

巴塞爾協定三

Basel III

黑池

Dark Pools

商業智能

Business Intelligence

RMBI

RMBI

RMBI

NE

WS

L

E

TTE

R

風險

風險

風險

ISSUE

02

FEBRUARY 2011

(2)

First of all, wish all of our readers a happy and prosperous year of the Rabbit!

In January 2011, HKUST organized an information session for secondary six students from early admission scheme. Many students were interested in RMBI Program and asked lots of questions about the program. One of the students approached me after the information session and told me that he would choose our program because he liked the idea of learning through practice and from others’ experience.

This is exactly one of our major principles when we designed the program. We endeavor to link our students to the business environment. Therefore, we always invite industry experts to share their experience in various areas in risk management (RM) and business intelligence (BI). In the past three months, we have already organized three seminars for our students; one seminar was about operational risk, one about BI software and the other about the career in risk consulting.

Moreover, we have launched a mentorship program for our students. Until now, we have had more than 35 mentors from various local and overseas’ companies. The mentorship program can help our students to look beyond the classroom and learn from the industry experts. I would like to take this opportunity to thank again for our mentors’ time and patience with our students.

In addition to learning from others, putting what have learnt into practice is important for our students to enhance their knowledge and demonstrate their ability in handling RM and BI issues. We have aligned some companies to provide capstone projects for our students. Each company first identifies a research problem in RM or BI and a group of 4-5 students, with the guidance of our university professors and the company’s representatives, will participate in solving the problem. Capstone projects for our first batch of students will start soon. If you are interested in this collaboration, please let us know.

In the issue, our students have interviewed experts from industry and academic. They discussed some risk management or business intelligence practices in their organizations and also the new Basel III regulations. The “digest” about dark pools brings new insights into the current hot financial topic. Students intend to include information about RM exams and BI definitions in this issue for readers’ interests too. Hope that you enjoy reading this newsletter. Prof. Mike So ([email protected]) 首先,恭祝各位讀者兔 年快樂、恭喜發財! 二零一一年一月,科大 為合資格參加拔尖計劃 的中六生舉行資訊日。 很多學生對本課程深感 興趣,並就課程提出大 量查詢。資訊日結束後 有一學生向我表示,他 很贊同本課程透過實踐 和經驗分享進行學習的 安排,並會報讀本課程。 這正是我們設計課程大 綱時的其中一個重要原 則。我們力求讓學生與 商業環境接軌。所以我 們經常邀請業界人士向 學生分享在風險管理或 商 業 智 能 上 的 工 作 經 驗。過去三個月我們已 舉行了三個講座,分別 圍繞營運風險、商業智 能 軟 件 及 風 險 顧 問 工 作。 另外,我們亦為學生籌 辦師友計劃。現在我們 已有超過三十五位來自 多間本地和海外公司的 導師。這計劃能幫助學 生接觸外界,向業界專 才學習。我想藉此機會 感謝各位導師為學生所 花的時間和耐性。 除 了 向 有 經 驗 人 士 學 習,活用所學以增進學 生掌握風險管理和商業 智能的能力及知識也非 常重要。我們已經聯絡 一些公司提供專題研究 報 告 題 目 予 本 課 程 學 生。每間公司會指定一 個風險管理或商業智能 的研究項目,在本校教 授和該公司代表的指引 下,讓四至五人的學生 小組解決問題。第一屆 學生的專題研究報告即 將展開。如你對這合作 計劃有興趣,歡迎和我 們聯絡。 本期的新聞報中,學生 訪問了來自學術界和業 界的專家,討論公司如 何應用風險管理或商業 智能,以及剛推出的巴 塞爾三協定。同時亦會 就「黑池交易」這議題 作出剖析。學生同時因 應讀者興趣,刊載風險 管理專業考試有關資料 及商業智能的解說。希 望您喜歡這份新聞報。 蘇家培教授 ([email protected])

3

Integrating Theory and Practice in RMBI Program

理論與實踐並重的課程

Risk Control with Business

Intelligence tools in Airline Industry

航空業風險管理及

商業智能工具的運用

7

Laying Off Credit Risk in Banks

銀行信貸風險管理

10

Discussion on Basel III

論巴塞爾協定三

12

Dark Pools in Hong Kong

黑池在港

15

Introduction to Business Intelligence

商業智能簡介

18

Well-recognized Risk Management

Certificates

20

Glossary

術語彙編

22

RMBI Program Development

RMBI 課程發展

contents 目錄

HKUST RMBI Newsletter Editoral Board

Advisors Prof. Mike So, Program Director of RMBI Dr Adela Lau, Senior Instrutor of RMBI

Editors Lam Hing Ki, RMBI year 2 Leung Ka Tai, RMBI year 2 Pang Ming Chuen, RMBI year 2 Wong Chun Ko, RMBI year 2 Wong Chung Yi, RMBI year 2 Yeung Wing, RMBI year 2 Yeung Yuk Ting, RMBI year 2 Lu Qin Yuan, QFIN year 1 Ng Man Hin, RMBI year 1 Tam Hiu Fung, RMBI year 1 Yet Pak Hei, RMBI year 1 Yim King Pong, RMBI year 1 Publication February 2011

Contact

Risk Management and Business Intelligence Program Office Phone 23585098

Email [email protected] Website http://www.rmbi.ust.hk

© 2011 Risk Management and Business Intelligence, HKUST

All rights reserved. Nothing may be reprinted in whole or in part without written permission from the publisher.

Image Courtesy

Kansai International Airport Co., Ltd (P.4-5 background) Zuzu (P.6 background)

(3)

風 險 管 理

及商業智能

工具的運用

Risk Control with Business

I n t e l l i g e n c e t o o l s

Interview with 訪問

Steve Tunstall

Head of

Corporate Risk Management

at Cathay Pacific Airways

國泰航空公司

企業風險管理主管

(4)

I: Interviewer

S: Steve Tunstall

I: What kinds of services does Cathay Pacific Airways provide in its daily operations? What are the corresponding operational risks?

S: Cathay Pacific is a large group with diverse array of services. It mainly comprises of three airlines – Cathay Pacific, Dragon Air and Air Hong Kong. In addition to the aircraft fleet, CX has investments in

nothing. For example during the extreme snow storm in London last December, we think we generally perform better than our competitors in handling the unexpected situation. We arranged buses to drive people from London to Manchester, and sent out three extra flights to fly them back to Hong Kong. But still I saw many people stunned in the London airport, yet the situation is difficult to improve in the meantime. We need to learn from these experiences to make sure we can do better next time.

I: What could possibly be done or have already been done in order to reduce the impacts caused by the weather problem in London?

S: The location of the problems is crucial. It affects our actions to be taken. For example, London and New York are two big cities with busy air traffic. There is hardly any slack for the flights. Therefore, when the schedules were disturbed last Christmas, the entire airline operations catering, cargo and laundry services.

Among all investments, airline business contains the most uncertainties. Our first priority is to ensure the safety of carrying people to the high atmosphere. But economic factors are also our major concerns. Whenever the economic cycle changes, any negative impact on the airline business will be magnified. This large effect is comparatively unusual in most of other industries.

I : I s w e a t h e r a l s o o n e o f t h e concerns? How does the recent weather problem in London last December affect your operations?

S: The weather affects our business a lot. Although weather is unpredictable, we must take contingency actions as fast as we can. We put a lot of time and energy to try to minimize its influence. Unfortunately, sometimes we can do

問:本報記者 答:Steve Tunstall 問: 國 泰 航 空 公 司 日 常 營 運 會 提 供 什 麼 服 務? 有 那 些營運風險牽涉在內? 答:國泰航空是一間規模龐大 而架構複雜的公司。我們企業 包含三間航空公司,包括國泰 航空、港龍航空及香港華民航 空。 除 了 我 們的航空團隊 以 外,國泰更投資於飲食、貨運 及洗衣服務。 在眾多的投資項目當中,航空 業存在較多的未知數。因為首 先我們要確保搭載乘客到高空 的安全。再者,經濟因素也是 我們的主要關注項目,每當經 濟周期轉變,任何對航空業務 的負面影響也會被放大。而這 影響於大部份其他行業中並不 常見。 問: 天 氣 變 化 是 否 其 中 一 個 顧 慮? 去 年 十 二 月 份 倫 敦 的 惡 劣 天 氣 對 你 們 國 泰 的業務有何影響? 答:天氣變化十分影響我們的 業務。即使天氣變幻莫測,我 們仍必須盡快採取應變行動。 我們投放很多時間和資源在這 一方面,以減低天氣對乘客的 影響。可是,有時候我們卻無 能為力。例如去年十二月份, 倫敦暴雪就是一次意料之外, 但我們認為於這樣的情況下,我們的反應已比其他航 空公司來得合適。我們安排巴士接載乘客,由倫敦機 場前往曼徹斯特,並派出三架加班航機從曼徹斯特接 載乘客回港。不過仍然有很多人滯留在倫敦機場,可 是這情況在當時實在難以改善。我們需從中學習,確 保下一次遇到同類情況時能做得更好。 問:有甚麼可預先準備或已準備來減低這些倫敦 惡劣天氣產生的影響嗎? 答:發生問題的地點是關鍵所在,可影響我們所採取 的行動。例如,倫敦及紐約是兩個航空交通非常繁忙 的城市,難以有任何閒置時間讓航機升降。故去年聖 誕發生的突發事故打亂航程安排時,整個航空營運也 受到影響,亦沒有即時的解決方法。比起其他較少客 運需求的機場,繁忙機場相對地會受到更嚴重的影 響。況且,倫敦於香港六千里之外,要確保資源分配 適當十分困難。不過如果類似情況 發生在我們的根據地香港的時候, 我們可以迅速而有效地應對各種情 況。但我們會繼續從經驗中學習掌 握不同的情況,以能夠靈活應變各 種意外為目標。 問:你們如何處理與恐怖襲擊相 關的風險呢? 答:我們十分關注恐怖襲擊問題。 譬如現在國泰的管理階層正在密切 注視朝鮮半島的局勢,但現時很難 預計當地的局勢發展。雖然我們沒 有航機直飛往北韓,但仍有飛越北 韓上空的航機。若繼續維持原定航 線,我們便需考慮其隱藏的危機。 此外,我們有很多航班由香港飛往 歐洲,大部分都會飛越過中東地區。 我們要審慎考慮這些航班的路線, 4

(5)

were affected, and there was no immediate solution t o t h a t . T h i s e f f e c t i s undoubtedly more dramatic than those airports with less demand.

Moreover, as London is 6 0 0 0 m i l e s a w a y f r o m Hong Kong, it was much more difficult to make sure resources are well allocated. B u t i f w e h a v e s i m i l a r situation happened in Hong Kong, our home base, we could respond much more effectively to make sure the situation is under control.

But certainly, as the last line of defense – air craft, we would also be well-prepared to stay alert to any difficulties that may happen.

Addressing the terrorist problem, what we encourage to do is thinking about long term trends. For example, we study where the threats occur currently and how they affect us now and in the future. We also try to forecast where threats might occur in the next few years.

I: Addressing the risks mentioned, how does business intelligence (BI) benefit Cathay Pacific in the way of identifying, managing and mitigating risks?

S: We do this in a more structured way. Each department and subsidiary regularly produces risk registers, which identify major risks that we’ll face in the business. Each individual team in the department identifies their corresponding hidden risks and reports to the central risk management department, where all the information is gathered for further investigation. So that is BI.

Externally, we look for changes that might affect us. Those may be dramatic changes like terrorism, but may also be slow changes in technology or demography, which would be equally significant. We have to watch those and make sure that our strategy responds accordingly.

We also value experiences a lot. People who have been here more than twenty years have gone through many different business cycles. They know from their Nevertheless, we learn from every experience

in different situations to make sure that we can do our best in handling all kinds of accidents.

I: How to handle risks addressing the terrorist problem?

S: The terrorist problem is something we are very conscious of. The CX management team is currently looking at the situation on the Korean Peninsula, where at the moment it is difficult to predict what might happen over there. While we don’t fly to North Korea, we do overfly North Korea. So we have to consider what might be the hidden risks. We also have a lot of flights to Europe from Hong Kong. Many of those go over the middle-east. We think thoroughly what our routine is through the middle-east to make sure we have got the right route towards there.

To effectively control the risks, we follow rules. Apart from the best practice laid down by the HK government, we take regulations of stringent authorities in America and Europe as our level of security. When we are on the ground, we closely follow guidelines given by the local government. For instance, we have an immigration watch list to identify any suspect. If there is any problem, we will notify the authorities at once. But most of all, airports are front lines to spot hidden terrorist problems. Most terrorism protection measures are taken by airport management.

確 保 我 們 經 過 的 路 線 是正確的。 為有效控制風險,我們 遵守規定。除了香港政 府訂定的安全指引,我 們 亦 同 時 跟 從 美 國 及 歐 洲 當 局 所 發 出 較 嚴 格的規例,並且視之為 我們的安全標準。在地 面 我 們 會 依 照 該 地 政 府指示。例如,我們擁 有 一 入 境 監 察 名 單 用 作識別任何可疑人物。如發現任何可疑,會立刻通知有關 當局。機場始終是第一防線,因此機場的管理層已做了很 多防範恐怖襲擊的措施,避免機場成為襲擊目標。當然作 為最後防線,航機應隨時作好準備,應付任何突發事件。 對於恐怖襲擊問題,我們一直鼓勵把目光放遠一點。例如, 多想想哪些地方會出現危機,這些危機出現時及在將來會 怎樣影響我們。我們還嘗試預測在將來數年哪些地方會受 到威脅。 問:商業智能如何有利國泰識別、管理和減輕上述風 險? 答:我們會有系統地進行。各部門和子公司會定時製作風 險登記冊,識別我們在業務中將要面對的主要風險。然後, 部門中的各小組會識別出與它們相關的穩藏危機,並向中 央風險管理部門報告,再將所有資料收集作進一步調查。 這就是商業智能。 對外,我們會找出可能影響我們業務的各種轉變。可能是急 劇得如恐怖主義的變化,亦可能是緩 慢得如科技或人口的轉變,對我們來 說同樣重要。我們要留意這些變化, 確保我們有相應的策略應對。 另外,我們亦非常重視經驗。在這裡 工作超過二十年的人,已經歷過許多 不同的商業週期。基於過去經驗,他 們可預知將會發生什麼事情。即使此 刻這些問題仍不算明顯,他們亦可洞 察潛在的危機並據此提出相應行動。 數據挖掘雖然重要,但人的智慧和經 驗亦不能忽視。 問: 國 泰 有 否 使 用 商 業 智 能 軟 件 或電腦軟件來處理數據? 答:我們有很多不同的電腦軟件處理 不同工作。但我們沒有一套既定的商 5

(6)

experiences what is going to happen. Even if the problem is insignificant at the moment, they can spot the underlying risks and suggest actions to be taken accordingly. So capturing the goodness from people’s brains is often much meaningful than simply data mining.

I: Does Cathay Pacific use BI software or computer software to handle the data?

S: We have a lot of different computer software doing different things. But we do not have anything so called "BI package". The reason is that we have many different pieces of software serving different purposes. It might be quite easy for small organizations to handle data with a single database. But that is pretty hard for complex organizations like CX. We are not storing data in one database, but in a wide number of different systems instead. I don't think there is a BI package in the market that could cater to all our different types of software.

customer details?

S : D a t a p r i v a c y p r o t e c t i o n i s a crucial issue that has to be treated appropriately. Otherwise accidental leakage of customer information will cause great damage to the reputation of the company.

We have set up committees looking at these issues for over a decade, though it only becomes prominent in the Hong Kong media last few years. Of course we can always do better and more to make sure that we don’t make any mistakes in this area. We are now improving systems to look at our strengths and areas for improvement. We also study cases and learn from others’ experiences. There have been some quite dramatic cases, not only in Hong Kong. Some of the data security bridges that I have seen overseas have been very significant for the company's concern.

We do use BI though, for example, we have very sophisticated software. While we auction passenger seats, this software helps to maximize our yields by looking closely to the c u r r e n t a u c t i o n p r i c e and predicted price. The prediction fundamentally depends on BI software to tell what the demand is at a particular time. Therefore, it is always better to have highly complex software to back up the analysis. But we first have to solve a lot of problems in order to have such software.

I: How to reduce the risk of the leakage of information, such as 業智能套件。原因是我們有許多不同的 軟件,用作不同用途。小型機構較為容 易用單一數據庫處理數據,但像我們這 樣業務繁複的組織則相當困難。我們不 是存放資料於單一數據庫,而是分散在 很多不同的系統上。我不認為市場上有 一個商業智能套件能迎合我們各種軟件 的功能。 但我們確實也會用到商業智能的概念, 例如我們有非常複雜的軟件。當我們拍 賣每個客機座位時,這些軟件就可以協 必須適當處理。否則,意外洩漏客人資 料可能導致公司聲譽嚴重受損。 過去數年,香港媒體才開始報道數據安 全問題,但事實上我們於十多年前已 成立委員會應付這些問題。當然,我們 還可以做得更好、更多,以確保我們在 這方面不犯任何錯誤。我們正在改善系 統,看看哪些部分是我們的優勢,哪些 部分仍待改進。我們亦會參考過往案例 及從其他公司的經驗中學習。在香港以 及其他地方已有一些較嚴重的數據安全 個案,而我考察過的部分海外數據傳輸 通道亦引起公司的注意。 助密切留意現時拍賣的價格和預測未來 價格的走勢,使我們能從中得到最大利 潤。基本上,由商業智能軟件得出在某 個特定時間的座位需求與預測的價格有 直接關係,所以如果能夠使用專門的軟 件作後援,預測就會更準確。但在這之 前,我們必須要解決很多問題,方能達 到該目標。 問:如何降低資料 ( 如客戶個人資料 ) 洩漏的風險? 答:我們非常重視數據私穩保護問題,

I am very lucky here that there is a very good understanding at the senior management level, so they have the risk team at the right level in the organization. But sometimes it is quite difficult for the risk voice to be heard and it is important that it doesn’t get subsumed.

When there are many people in the organization, we would say, “Oh, here is opportunity to make a lot of money. Let’s get in there fast.” But the role of the risk management is not to slow that down; we also want to understand what risks are taken. So let’s not worship blindly to these opportunities but say, “What might go wrong? Let’s think about that thoroughly.”

 能夠在這裡工作,我感到 很幸運。管理層了解風險管 理的重要,使風險管理團隊 擁有適當地位。但有時與風 險相關的意見卻難以傳達至 管理層並被採納。 當組織中有人熱烈追求更多 盈利時,風險管理人員的角 色不是影響其進程。我們是 想了解這樣做要承受什麼風 險。所以我們不盲目地接受 這些機會,而是對這些機會 作徹底分析。

Mr. Tunstall's words

Steve Tunstall 的話

6

(7)

US Subprime Mortgage followed by European Union Debt Crisis had put an alarm to credit risk management in banks. In this session, based on interviews with two industry experts, we shall provide an overview of the loan lending procedures in banks to foreign companies, and deliver an insight about how banks are staying away from different kinds of exposed risks within the business.

The first interviewee is a MBA and CFA holder with over 22 years of banking experience in leading international banks in diversified banking areas including credit risk management, corporate and commercial banking relationship management, global transaction banking, direct investment and corporate finance. The second interviewee is currently vice president of a local risk management educational organization and a Certified Risk Planner. With great banking experience in corporate lending, she has been giving lectures in financial risk management at the institute and other associations.

General Process

Generally, banks would not lend to foreign companies directly unless the foreign counterparties have been well investigated. Most likely the loans are made to local company as the bank may have better knowledge in understanding the local business trend rather than in other countries. These companies must have certificate of incorporation specifically for foreign companies and business

registration showing their registration in Hong Kong. A lawyer is also hired for research on the company’s reputation in its own country, and proving with a certificate of good standing, which shows detailed information of the company such that lender can have good justification of credibility of the company. As a

common bank practice, credit relationship managers of lending bank will monitor the business and financial performance of the borrowers via regular customer calls with the management / owner of the company and site visits if applicable, as well as via updated financial and audited reports of the company.

Managing the risks

When lending to a foreign company, bank should pay attention to two specific details: i) where the company is registered; ii) where their assets locate, as assets in countries with politic instability will subject to political turbulences frequently. There

美國次按危機及緊隨的歐盟 債務危機是銀行信貸風險管 理的一個警報。本節綜合與 兩位業界人士進行之訪問, 概述本港銀行向外國公司貸 款的程序,並剖析銀行如何 規避業務中面臨的不同風 險。 第一名受訪者擁有工商管理 碩士學位及特許財務分析師 資格,在數間大型銀行工作 合共超過二十二年。曾任職 的部門包括信貸風險管理、 公司及商業銀行服務管理、 環球交易、投資與公司財 務。 第二名受訪者則為本地風險 管理培訓團體的副主席,同 時擁有公認風險策劃師資 格,在銀行企業貸款這範疇 極具經驗。此受訪者亦為其 所屬和其他組織提供培訓服 務。

一般程序

除非已對海外公司進行深入 調查,否則銀行一般不會向 它們進行借貸。相反,銀行 會貸款給該公司在香港註冊 的子公司或分公司。這些公 司必須有外國註冊證書和其 在香港登記的商業證明。銀 行還會聘請律師審查該公司 在其國家的信譽,發出證明 該公司有良好信譽的證書, 並提供該公司的詳細資料, 以便放債方正確評估公司信 用的狀況。根據慣例,銀行 的信用客戶經理會透過定期 聯絡客戶公司管理層或公司 持有人,及閱讀公司最新的 財務與審計報告,監察其營 商和財政表現。

風險管理

當貸款給一家外國公司, 有兩點是銀行必須注意的: i) 公司的註冊地和 ii) 其資 產所在,因為位處政局動盪 國家的資產經常受動亂威 脅。此外,自然災害和匯率 波動可能影響資產價格。為 減少不同國家政府帶來的風 險,銀行制定信用風險評級

Laying off

Credit Risk

in Banks

銀行信貸

風險管理

7

(8)

may be influences on assets’ value due to natural disasters and exchange rate fluctuation. Therefore, in order to reduce problems caused by variance in different countries, banks will set up risk rating policy for credit risk corresponding to their location. By introducing swap contracts, bank can even reduce 80-90% of related loss.

Credit risk rating policy

Most banks have specific credit policies to rate credibility of borrowers. Below are some usual factors that affect the probability of default (PD) of a company. However, changing in only one factor may not necessary enough to have revision in the PD.

(i) Financial performance of borrowing company –

By digging deeply into the financial statements of borrower and looking into financial ratios and abnormalities, banks can understand the business trend. Banks can also check if there is any high variation in revenue or changes in debt pattern and component, so as to make sure they are lending to a company which is able to repay its indebtedness.

(ii) Company management –

Exploring major management problems existing within the company and

investigating the company’s integrity and data disclosure policies.

(iii) Future of the industry –

By discovering what is the nature of the industry and the company’s position in respective industry, analyzing whether the company is in promising or deteriorating industry. Banks can make a more precise judgment on whether they should lend out the loan.

(iv) Respective country’s information –

By learning about the country’s

development, laws and regulations, banks can estimate how sensitive is the company towards the fluctuation in interest rate and currency exchange rate, and determine the risk exposure of its own.

Based on above information, banks then set up credit risk measuring model. Although the process is costly and time consuming, it is a more reliable and effective method comparing to external agencies’ ratings. Credit risk measurement involves

calculation on expected loss (EL), which is a function of the PD and loss given default (LGD) as well as other relevant parameters. 政策時亦會考慮這些因素。 通過購買掉期合約,銀行甚 至可以減少 80-90%的相關 虧損。

信貨評級機制

大多數銀行都會有特定的信 貸評審政策。以下是影響貸 款違約率的一些主要因素。 不過,任一因素改變未必足 以對貸款違約率構成影響。 (一)借款公司的財務業 績— 透過分析公司的財務報表和 研究其財務比率和異常,銀 行可預測其業務趨勢。此 外,銀行也可核查該公司的 利潤波幅、債務結構和未來 變動,以確定公司有能力償 還負債。 (二)企業管理— 檢測企業現存的管理問題, 查察他們處理道德操守和數 據發佈的政策。 (三)行業前景— 了解行業性質及該公司在同 業中的表現,銀行可得知其 所屬行業是前景明朗,還是 夕陽工業,從而得出更有力 的證據決定是否批出貸款。 (四)國家資訊— 透過了解國家發展、法律與 法規,銀行可估計該公司對 利率及貨幣匯率波動的敏感 度,有助銀行估計放款的風 險。 根據以上的資料,銀行會建 立信貸風險測量模型。雖然 過程會花費大量資源和時 間,但這方法卻較信貸評級 機構所給予的評級更為可靠 和有效。 信貸風險測量牽涉預期損失 (Expected loss) 的計算。 預期損失是貸款違約率 ( 違 約率 ) 、違約損失率(Loss given default)( 損失率 ) 和其他變數的函數。前兩者 可使用信貸評級模型得出。 對於風險較高,即較高違約 率的公司,銀行慣例是向它 8

(9)

These two can be estimated by credit rating models. For companies with higher risks (i.e. higher PD), common practice among banks is to ask them for more collateral, or to establish suitable repayment structure to minimize LGD. In the ideal situation, bank shall aim for borrower with lowest PD and also lowest LGD by obtaining as much collateral as possible. However, in practice, companies with low PD (i.e. good credit rating) will require banks to lend on unsecured basis. Due to fierce competition in the bank market, stronger credit rated borrowers generally can command a lower loan pricing and also with less stringent terms and conditions, and loan covenants.

Credit Rating Agency

As mentioned above, large commercial banks usually have their own risk analysis model and rely less on credit rating agencies. However, trading house and investment bank do have higher reliance on credit rating agencies.

Recently, banks start to blame the agencies for not assessing the rating accurately. One of the cases is that they failed to degrade deteriorating companies timely such as Lehman Brothers, which got a rating AA+ before collapsing. In addition, conflict of interest is a matter as rating agencies always acting as paid consultants, give suggestions for companies on how to maintain certain ratings.

Remarks

Loans can be structured on a committed basis or on an uncommitted basis

depending on the needs of the borrowers as well as their credit quality. In general, banks will consider committed loans to stronger credit rated borrowers. For committed loans, bank is committed to lend and cannot demand repayment as the bank likes unless there is an event of default by the borrower. For uncommitted loans, the bank reserves the rights to lend and has the rights to demand repayment at its own discretion.

It is bank’s responsibility to performance regular credit review on the borrower. If a bank identifies signs of weaknesses or problems about the company after lending out the money, e.g. downgrading by credit rating agency, there is certainly a list of measures to be carried out. However even a bank has carried all the measures before hand, it does not mean the bank is totally free from risk.

們取得更多抵押品,或設立 適當的還款安排以降低損失 率。理想情況下,銀行會索 取最大量的抵押品,把違約 率及損失率減至最低。然而 實際上,低違約率的公司會 要求銀行以無抵押形式批出 貸款。另外由於銀行業競爭 激烈,信貸評級較高的客戶 一般享有較低借貸利率與更 寬鬆的條件和規定,以及貸 款協議。

信貸評級機構

如上文所述,大型商業銀行 通常有自己的風險分析模 型,所以較少依靠信貸評級 機構。然而,交易公司和投 資銀行對信貸評級機構有較 高的依賴。 銀行近來開始指責信貸評估 機構給予的評級不準確。其 中一個例子是,他們未能及 時把財務狀況惡化的公司降 級,如倒閉之前獲評級為 AA + 的雷曼兄弟。另一個 問題是利益衝突,因為評級 機構會受雇於企業,以顧問 的角色就如何保持某信貸評 級提供建議。

備註

視乎客戶需要和它們的信貸 紀錄,銀行會以有責任或無 責任兩種形式進行放款。銀 行多數會考慮向擁有良好信 貸評級的公司借出責任性貸 款。責任性貸款的情況下, 銀行有義務借出貸款。同時 若欠債公司沒有任何違約情 況,銀行不可索求提早還 款。在無責任的情況下,銀 行保留借款權利,亦有權酌 情要求客戶提早歸還借款。 定期檢查借貸公司的財務狀 況是銀行的責任。銀行如果 在批出貸款後發現如借貸公 司信貸評級下降等問題,便 需採取一連串應對方案。不 過即使銀行放貸前實行一切 措施,仍未能擺脫所有風 險。 9

(10)

Discussion on

Basel III

In

2008, the collapse of Lehman Brothers was the prelude of the most significant world event in the past few years – the financial tsunami. This collapse caused a chain effect in the banking sector. Banks ran out of capital and they needed to raise funds in order to fulfill the regulations conducted by central banks. The credit market contracted immediately and investors lost confidence at that time. To prevent this situation from happening again, Basel committee launched Basel III to tighten the capital requirement of banks. However, the banking industry faced dilemmas. On one hand, banks obviously do not want tsunami come back. On the other hand, they worry about adopting new Basel regulations as this may affect their revenues. To know the exact effect of Basel III, we need to have an in-depth examination on the content of Basel III.

Basel is a regulation record proposed by Basel Committee on Banking Regulations and Supervisory Practices, which consisted of representatives of central banks from about 20 major countries, including China and Hong Kong. After the implementation of Basel, all participating countries shall establish relevant laws in accordance with Basel to monitor the banking industry. However, Professor Peng Xian Hua added that the nature of Basel is only a recommendation and has no legal force. As Basel II did not consider impacts from undesirable economic condition and risks such as tail risk; the capital requirement before financial tsunami was totally insufficient to confront the tremendous losses incurred from financial risks. In 2008, financial tsunami occurred around the globe and loopholes of Basel II were revealed. Thus the introduction of Basel III was made based on mending the leaks in Basel II. Such modifications include the increase in capital ratio and Tier 1capital ratio. Tier 1 capital is the core measure on the financial strength of banks and it mainly consists of common equity. During the financial tsunami, many banks claimed to have considerable amount of Tier 1 capital, but in fact, a significant portion of their Tier 1 capital were far less loss-absorbing than common equity, which resulted in the loss of confidence by investors towards the Tier 1 capital ratio. The significant amendment by Basel III has clearly and stringently defined capital and Tier 1 capital as well as which aims to enhance the quality of bank capital. For example, some less loss-absorbing capitals cannot be counted as Tier 1 capital anymore, which in turn increased the

loss-absorbing ability and loss-resistance ability of banks in respect of Tier 1 capital. Thus, a more stringent regulation on bank regulatory capital can significantly increase investors’ and depositors’ confidence in banks, enhancing their investments and deposits in banks. For banks, one of the obvious advantages is that their capital can serve as a cushion in worsening financial situation. Professor Peng pointed out that the financial crisis in 2008 witnessed that regulatory capitals were significantly underestimated. The actual losses of banks during the financial tsunami were far above the regulatory capitals calculated in Basel II. This illustrated that tail risks were crucial in estimating regulatory capitals that banks should hold and Basel II neglected imposing capital requirements for some key types of risk such as counterparty risk. Thus the increase in regulatory capital is essential in absorbing losses in undesirable situations. ‘This leads to another advantage.’ Professor Peng suggested. Because of higher requirement in regulatory capital, credit rating of banks may be upgraded and hence they can take the edge over it by offering lower bond yield and getting better terms when they issue new bonds and stock shares. This may lower their cost of funding in the long run.

In the short run, the stricter requirement in Basel III will pile pressure on banks in bearing higher cost of funding. Banks may transfer the costs to other parties such as borrowers by increasing interest rates. This would induce additional investment costs and investors will therefore be more conservative

in making investment decisions. Investment activities may become less aggressive in the way that investors are not willing to pay high costs. Basel Committee and IMF (International Monetary Fund) projects that the transition to higher capital standard will have a modest impact on aggregate output. Every one percent increase in common equity to risk-weighted-asset ratio will reduce annual growth rate by 0.04 percent over a four year period, mainly due to slow down of investment. In addition, impacts to banks also differ from region to region. For example,

banks in the US will likely have more difficulty in raising high level of their own common equity than those in China, while banks in China will be more challenged by requirement on advanced risk management system. Investing in simpler asset classes, banks in China need to allocate more resources in developing a more structured risk management system and this will incur costs.

Despite some possible negative effects, Basel III Accord can bring significant influences that can benefit the whole world. Professor Peng believed that it not only increases the stability of the financial system, but also promotes sustainable growth of the economy, by means of reducing the probability of financial crisis. If financial crisis happens, the aggregate output of a nation will decrease and thus decrease the employment rate and harm the whole society. So the stability brought about by the new Basel accord can reduce the GDP volatility and help to avoid a crisis. What’s more, a healthier and safer banking system can also absorb losses from even outside of it, thus protecting the mainstream economy.

In a word, although the new Basel accord could decrease the investment, which in turn, leads to a very limited decrease in GDP, it can introduce much more stability and ensure a sustainable growth of the whole economy. ‘It is not difficult to see that the new accord’s great benefit can well offset the minor decrease in the growth of economy, and is worthwhile to be applied all around the world.’ Professor Peng concluded.

Peng Xian Hua

Assistant Professor, Department of Mathematics, HKUST

(11)

巴塞爾協定三

彭獻華

科大數學系助理教授

零零八年,雷曼兄弟的倒閉為過去數年的世界大事—金融海嘯揭開序幕。 這次倒閉於銀行界引發連鎖反應,世界各地多間銀行面臨財政危機,需要透 過集資來滿足中央銀行對銀行資本的要求。另一方面,信貸市場不斷收縮,投資 者對市場失去信心。為防止這種情況再次出現,巴塞爾銀行監管委員會推出巴塞 爾資本協定三 (Basel III),從而收緊對銀行資本的要求。可是,銀行界卻面對兩難 局面。一方面,業界不希望金融海嘯重臨。同時它們對新巴塞爾協定三的推行存 在憂慮,因為這或會影響銀行盈利。要了解新巴塞爾協定三帶來的真正影響,我 們必須深入探究巴塞爾協定三的內容。 巴塞爾協定是一項由約二十個國家的中 央銀行代表組成之巴塞爾銀行監管委員 會所提出的協議,其委員會的成員國包 括中國及香港。在巴塞爾協定推行後, 成員國需設立相關條例從而監管銀行 業。不過彭獻華教授補充,巴塞爾協定 只是建議性質,且並無任何法律效力。 因為巴塞爾協定二沒有考慮未能預測 的經濟環境及風險,如尾部風險 (Tail Risk),所以在金融海嘯前的資本要求 根本不足以去面對金融危機中出現的巨 額損失。二零零八年金融海嘯席捲全 球,舊巴塞爾協定漏洞浮現。因此,巴 塞爾協定三被加快推出以彌補這些漏 洞。巴塞爾協定三是基於巴塞爾協定二 進行修訂所得。修訂包括增加銀行資本 比率及一級資本比率。一級資本是指對 銀行財務能力一個核心衡量,而當中大 部份是普通股本。很多銀行在金融海嘯 時聲稱擁有很多一級資本,但實際上相 當大比例的一級資本吸收虧損的能力遠 低於普通股本,這就導致投資者對一級 資本比率失去信心。所以巴塞爾協定三 所做的一個重要的修訂是更清楚和嚴格 的定義了資本和一級資本和提高了資本 的質量。例如,一些對虧損的吸收能力 差的資本不可以算作一級資本,從而增 加了一級資本對虧損的吸收能力和銀行 承受損失的能力。所以更嚴謹的受監管 資本規定可以大幅提高投資者和存款者 的信心,從而增加他們於銀行的投資和 存款。 對銀行業界而言,一個顯而易見的好處 是其資本可在財務狀況惡劣的時候作為 緩衝。彭教授指出二零零八年的金融危 機便證明銀行需要具備的受監管資本被 嚴重低估:金融危機中銀行的實際損失 遠比舊巴塞爾協定所定明的監管資本要 高。這表明尾部風險在計算銀行所承受 的風險中極其重要,並且舊巴塞爾協定 沒有要求銀行針對一些重要的風險持有 監管資本,例如交易對手風險。因此增 加受監管資本以吸收未能預測的虧損是 必須的。「這衍生另一好處。」彭教授 說。基於更高的受監管資本要求,銀行 的信貸評級可能會被提高,這樣銀行能 夠以更低的收益率發行債券和更高的價 格和條件發行新股,因此降低他們的借 貸和融資成本。因此,長遠來看,巴塞 爾協定三或會有助於降低銀行的融資成 本。 近期來看,由於巴塞爾協定三向銀行施 壓須達到較高的資本要求,這可能會導 致銀行提高借貸利率,將融資成本轉加 給其他各方如借貸者。對投資者來說, 因為銀行的借貸利率提高意味著額外的 投資成本,而他們又不願負擔高昂的費 用,所以投資決定會趨向保守,導致經 濟活動產出減少。但是,根據巴塞爾銀 行監管委員會及國際貨幣基金組織預 期,提高資本標準的過渡期對經濟產出 帶來的負面影響並不大。預期普通股本 對風險加權資産的比率每提高 1%, 在 未來四年內年增長率會下跌 0.04%,主 要由於投資放緩所致。此外,新協定對 銀行的影響會隨不同地區而有所改變。 以美國銀行為例,當它們提高自身普通 股本時,它們遇到的障礙會較中國的銀 行為多。中國的銀行則會於達至更先進 的風險管理系統方面遇上更多困難。故 此,投資於較簡單的資產類別時,中國 的銀行需要分配較多資源發展更具結構 性的風險管理系統,而這將會增加營運 成本。 儘管巴塞爾協定三可能會帶來一些負面 影響,但是同時可以為全球經濟帶來益 處。彭教授相信協定不但增加金融體系 的穩定性,而且通過減低發生金融危機 的機會率,可促進世界經濟持續增長。 如果發生金融危機,一個國家的生產總 值和就業率會下跌,整個社會將受到影 響。故新巴塞爾協定可以減低國民生產 總值的波動及避免產生危機。更重要的 是,一套更健康、安全的銀行系統甚至 可以吸收系統以外的損失,以保障主流 經濟。 總括而言,雖然新巴塞爾協定可能減少 投資活動,但對國民生產總值的下降是 非常有限,反而可以增加穩定性及確保 整體經濟持續增長。彭教授總結:「新 協定帶來的好處明顯地能蓋過帶來的壞 處,而且值得在世界各地推行。」

the city Basel, in which the Basel Accords are named after 巴塞爾,巴塞爾協定即以此城命名

(12)

in Hong Kong

在港

Dark Pools

黑池

According to the World Federation of Exchanges,

a total of US$18.911 trillion worth of shares

were traded on the electronic order books of the

exchanges in the Asia-Pacific region in the year of

2010, increased by 5.1% from 2009. That may imply

that the dark pools are growing in the region.

根據國際證券交易所聯會的數據,亞太區於二零一零

年總共有市值 18.911 萬億美元的股票透過電子交易

平台交易,較 2009 年增加 5.1%。這暗示黑池交易平

台正在亞太區進行擴張。

Dark pools are alternative trading systems and electronic communication networks, which are developed by big brokerages like Goldman Sachs. They allow investors (usually financial institutions like funds) to trade large blocks of shares outside traditional exchanges without public disclosure of the investors' identities, the trading volume and price, and regulators have limited visibility into those trades. They are now very popular in Europe and the US. According to Rosenblatt, around 13% of the trading volume in the U.S in December last year is from dark pools.

黑池交易平台是由一些大型經紀公司(例如高盛)開發的另類交易系統及電子通 訊網絡。投資者 ( 一般為機構投資者如基金等 ) 透過黑池交易,可以在交易所外 進行大手交易,而交易的詳細資料如他們的身份、交易量和成交的價格等,都不 會向公眾作出披露,透明度甚低,而監管機構亦只有有限的資料。黑池平台在歐 美地區深受歡迎。根據 Rosenblatt 的資料顯示,美國去年有 13% 的交易都是經 由黑池平台完成的。

What are dark pools?

什麼是黑池?

(13)

由於黑池交易平台能以匿名交易,機構投資者的資料及投資策略所承受的披露風 險較低,故此其較大的交易在市場上的影響亦相對減少。黑池交易平台亦能減低 交易的成本,並提供較高的交易速度。

Dark pools have made a relatively slow progress in Asia because the traditional exchanges are more protected by local regulations and there is no unified regulatory driver in the region. However, we do see some growth in Asia. 由於亞洲的傳統交易所較受當地監管機構保護,加上沒有統一的監管措施來推動 區內黑池發展,所以黑池交易平台於亞洲的發展較為緩慢。不過,我們仍然能夠 在亞洲看到黑池平台的擴張。

Dark pools in Asia

亞洲的黑池

Chi-East, the joint venture between the Singapore Exchange (SGX) and Chi-X Global, commenced their operations of dark pools on 11/11 last year. It is the first Asian traditional stock exchange-backed dark pool, now supporting the non-displayed trading of Singapore, Hong Kong, Japanese and Australian shares. The venture is regarded as an aggressive act for Singapore in terms of spurring competition between exchanges in the region and becoming a one-stop-shop for electronic trading in Asia. More importantly, it has threatened other Asian exchanges running monopolies or semi-monopolies, signaling them that they may need to face some competitions in the future.

去年十一月十一日,由新加坡交易所及 Chi-X Global 成立的合營企業 Chi-East 正式開始提供黑池平台服務。該黑池為亞洲區內首個有官方交易所支持營運的黑 池平台,它現時能夠讓投資者以匿名交易的方式買賣新加坡、香港、日本及澳洲 的股票。有意見認為新加坡此舉將會有助爭取其成為亞洲區內的一站式電子交易 中心,並且會引起區內交易所之間的競爭。更重要的是,此舉會對壟斷式或半壟 斷式經營的亞洲交易構成威脅,令它們今後面臨更大競爭。

In Hong Kong, there are about a dozen of dark pools. According to the Securities and Futures Commission, off-exchange transactions, including dark-pool ones and those using cross-trading platforms, reported to the Hong Kong Stock Exchange (HKEx) accounted for only about 3% of total market turnover on most trading days but the amount could exceed 10% on a few occasions like index rebalance.

香港現時大約有十餘個黑池交易平台。根據香港證監會的數據,在大多數交易日, 呈報香港聯合交易所的場外交易(包括黑池及使用交叉交易平台)所涉的交易 額,只佔市場成交總額約 3%。然而,在一些情況如指數重整時,總額可能會超 逾 10%。

Development of dark pools

in Hong Kong

黑池在香港的發展

Dark pools are not very popular in Hong Kong partly because of stamp duty, which amounts to 0.1% of the total value of the transactions. The absence of stamp duty in the United States reduces the transaction cost of trading and thus dark pools there can easily take the business away from traditional exchanges.

黑池在香港並不太普及,其中一個原因是佔交易總額 0.1% 的股票印花稅。美國 並沒有股票印花稅,故此交易的成本較低,而黑池就能夠藉此輕易地從傳統交易 所手中搶去生意。

The pools are useful for institutional investors since they allow investors to place anonymous orders, meaning disclosure of information and investment strategies are less likely to happen and thus the market impact of their big orders being reduced. Dark pools also help reduce transaction costs and provide higher trading speed.

(14)

根據香港法例,黑池必須取得自動化交易服務的牌照 (ATS)。如有買賣港股的服 務,它們必須為港交所的交易所參與者 (EP) ,並將交易紀錄匯報港交所。以往, 所匯報的交易會連同其他場外的交易會由港交所向公眾披露,但並無特別標記 那些交易是經黑池完成。由 1/2/2011 起,所有黑池的交易將加上「Alternative Liquidity Pool」(ALP) 的標記,用以將這些交易紀錄從其他場外交易紀錄分別出 來,以便港交所及證監會監察黑池交易。然而,港交所並不會對公眾披露所收集 到的交易資料,所以黑池對於公眾而言仍是神秘的,沒有透明度。

Striking the balance

爭取平衡

Regulators and exchanges in Asia are facing a tough decision on how they should react to the development of dark pools and the competition brought. Singapore has taken a positive stance in thinking that the development of dark pools could help itself become an international financial center in Asia as SGX has already had its joint venture with the Chi-X Global to set up Chi-East. Hong Kong seems to have a conservative view on the development of the pools. 亞洲的監管機構及交易所正需要作出一個困難的決定:它們該如何應對黑池的發 展及其帶來的競爭。新加坡樂觀認為發展黑池能夠幫助其成為亞洲的國際金融中 心,新加坡交易所已與 Chi-X Global 成為合營企業,成立 Chi-East。另一邊廂, 香港卻似乎對於黑池的發展甚為保守。

In an interview with South China Morning post (“Stock tax too useful to dump, Chan says” on 13/12/2010), the Secretary for Financial Services and the Treasury, Professor Chan Ka-keung, said that dark pools should be subject to more regulation. He believes stamp duty can prohibit dark pools to compete with traditional exchanges and Hong Kong should not allow dark pools to compete directly with the HKEx. It is clear that Hong Kong government does not want quick expansion of the pools and it concerns very much on the influence brought by the pools. Thus, it is highly possible that dark pools in Hong Kong will face more rules before they can gain larger market share in Hong Kong.

現任香港財經及庫務局局長陳家強在接受南華早報的訪問時 (“Stock tax too useful to dump, Chan says”, 13/12/2010),表示黑池應該面對更多的監管。 他認為股票印花稅能夠阻止黑池跟傳統交易所競爭,而香港應該避免讓黑池跟港 交所直接競爭。明顯地,香港政府並不想黑池在香港快速發展,亦非常關注黑池 帶來的影響。由此估計,黑池在香港獲得更大的市場份額前,其發展將會面臨更 多的監管。

The introduction of more yuan products in Hong Kong is sometimes regarded as the first priority of financial development in the city. Though not being the first priority, dark pools are getting more attention and the use of dark pools is a growing trend. Instead of combating the pools, regulators in Hong Kong should think of how to strike a balance between regulation and gaining market share. A dark pool backed or supported by the HKEx may be a good solution. It is not just appealing to regulators but also users because they may be suspicious about who are their counter-parties in the trade especially when using dark pools operated by banks. This, however, to Hong Kong, a city struggling with its trading hours and introducing more yuan products, may be too far to consider.

有人認為引入更多人民幣投資產品到香港應作為城內金融發展的首要項目。而黑 池雖然不是最優先的一環,但黑池漸受關注,交易亦是大勢所趨。香港監管機構 不應對抗黑池在本地的發展,而需要考慮如何在監管和爭取市場份額兩者中取得 平衡。一個港交所有份支持營運的黑池或會是不錯的選擇,它深受監管機構及使 用者歡迎,因為投資者在使用由銀行經營的黑池時會對交易對手的身份感到懷 疑。然而,這對於仍然在考慮延長交易時段和提供更多人民幣投資產品的香港而 言,似乎太長遠了。

Dark pools must be qualified as Automated Trading Services (ATS) under the law of Hong Kong. If they provide trading service of Hong Kong stocks, they are also required to be the Exchange Participants (EP). They are required to report their trades to the HKEx. Their trades were disclosed by the HKEx together with other off-exchange trades and there was no special identifier for those trades previously. Starting on 1/2/2011, dark pools transactions are labelled as “Alternative Liquidity Pool” (ALP) and be segregated from other off-market transactions and the activities of other alternative trading venues (ATV), allowing the HKEx and the SFC to have better tracks on the pools. However, the information collected by the regulators will not be disclosed to the public, meaning that dark pools are still dark to the public, without transparency.

(15)

Business Intelligence (BI) means the processes of analyzing company data in database, obtaining useful information and then applying analytic result on business decisions. This concept was suggested as early as thirty years ago, people start notifying its importance in recent years. Since there is an increase in computing capability of computer and a decline in price of data storage devices, enterprises are more willing to make use of BI. Information mined from data will help enterprises achieve different purposes, such as adjusting business operations to save cost, looking for potential customers as well as discovering unknown risks. Therefore, in the modern business environment which changes fiercely, BI is very useful. Here are two examples: 商業智能是指透過分析公司數據庫中的 數據,得出有用的資訊,並將分析結果 運用在商業決策之上。這個概念早在 三十多年前便有人提出,但到近數年才 為人所注視。隨着電腦的運算速度不斷 提高,而數據儲存裝置價格下降,刺激 企業使用商業智能。分析所得的眾多資 訊,可以協助企業達到不同的目的,如 調整營運方針減省成本、尋找潛藏客戶, 以及發現未知的風險等。所以在現代多 變的營商環境中,商業智能用途很廣。 以下是兩個商業例子:

1. Determining promotion targets

Many companies send leaflets or e-mails to existing customers, in order to increase customer spending. Some of them send promotion materials to all customers. In addition to making lots of waste, this promotion strategy is annoying to customers. Application of BI by analyzing customer records and their personal information can solve this problem by classifying customers and inferring their interest to different types of goods or services. Promotion target can then be easily determined. This greatly enhances efficiency of promotion campaign. 一、幫助決定宣傳對象 不少企業會向現有客戶寄送宣傳單張或發 送電郵,以增加客戶消費。部份公司會以 漁翁撒網的方式發送宣傳品給所有客戶。 這宣傳方式除了做成大量浪費外,亦對客 戶造生滋擾。應用商業智能分析客戶的消 費紀錄與個人資料,把他們分類,推斷他 們有興趣的貨品或服務種類,然後集中向 目標客戶進行推廣。這將便能大大提高宣 傳的效率。 2. Retaining customers

Some industries such as web hosting, pay television and especially, telecommunication have high customer churn rate owing to similar services offered by service providers and fierce competition. Loss of customers is considered as risk, because when compared with existing customers, finding new customers involves higher costs to make the same profit. Similar to previous example, customer information and their service using record can be used for BI purpose, estimating the probability of churn. Once these customers are identified, service providers may provide some discounts or contracts which better meet their needs, to let them become loyal customers. 二、挽留客戶 部份行業如網頁寄存、收費電視和特別是 電訊業,顧客流失率較高。這是由於服務 供應商提供的服務性質相似,而且競爭激 烈引致的。顧客流失被視為一種風險,因 為與現有客戶比較時,找尋新客戶需要牽 涉更高成本,卻只能賺取相同的利潤。正 如上例,客戶資料和服務使用紀錄可協助 商業智能分析,估計流失的機率。一旦發 現這些顧客,服務供應商會向他們提供優 惠或更切合其需要的合約,務求讓他們成 為忠實客戶。

I

ntr

od

u

ct

ion t

o

Bu

sin

e

ss

I

nt

el

lig

en

c

e

智能

簡介

15

(16)

Data Management

Amount of data is much more than what we imagine. For example, there are one million transaction records a year in a small supermarket. While a highway network in mainland involving less than ten entrances has records of three billion annually! To analyze data, we must first establish a database and data management system to access and process data.

數據管理 數據數量比想像中多。譬如一間小規模的 超級市場,每年交易數目可達一百萬宗。 而國內一個涉及少於十個出入口的高速公 路網絡,每年紀錄則是三億宗!所以要對 數據進行分析,必先建立數據庫和數據管 理制度,以便存取及處理數據。

Raw data should be processed and screened. This is because the amount of data can be very large, leading to burden on database. It is possible for errors to occur in data value. These values may be due to deliberate concealment or careless mistake by customers, data transmission problem or input error. Database administrators need to carefully consider possibility that the irrational numbers are wrong. If these values were judged as errors, causes of errors should be found to help with data correction. If correct values cannot be traced, wrong data may be discarded. Another approach is replacing error values with other values. Substituting with mean value from correct data is a common option. Another usual approach is estimating values according to correct data by maximizing its likelihood.

原數據亦應進行處理和篩選。除了因為數據量 可能非常大,對數據庫做成負擔外,數據中或 有錯誤數值出現。這些數值可能是由於顧客故 意隱瞞或無意中給予錯誤數字、數據傳輸出現 問題或有關人員輸入錯誤所致。數據庫管理者 需要仔細考慮不合理數值是錯誤的可能性。若 這些數值被判斷為錯誤,則要查找原因以便更 正數據。如果無法追溯正確數值,錯誤的數據 可能會被棄用,而另一方法是用其他數值代替 錯誤值。從其他正確數據取得的平均值是常用 的方法。此外借鑒其他正確數據,以概率最大 化 (maximum likelihood) 的方法評估數值亦 為常用手法之一。

Different input format is also a problem when handling data. For instance, date may be entered in form of year/month/day or year/ day/month. This uncertainty can result in confusion. So there should be clear guideline notifying staff to enter data in a specific way. 處理數據時常見的問題還有輸入數據時使用 不同格式。例如日期可按年月日或年日月的 格式輸入。這種不確定性可做成混亂。因此 處理數據時應有明確指引,讓員工以特定方 式輸入數據。

Different Analysis Algorithms

Many people who are not familiar with BI may think analytic algorithms are unfathomable or abstract. In fact lots of analytic algorithms are developed from mathematics. Following are three common analysis algorithms and the principles behind them.

不同的分析算法 很多不熟悉商業智能的人以為分析數據的算 法都是深不可測或十分抽象。其實不然。很 多分析算法事實上是基於數學原理。以下會 介紹三種常見的分析算法和其背後的原理。 1. Clustering

A data point may contain many attributes. Raise a financial company as an example. Its customer data records consist of customer’s age, occupation, salary, loan records and so on. Clustering can assign data into different groups according to their inter-distances. If there are only two attributes in a data, way of calculating data distance is similar to calculating distance between two points on a two-dimensional plane. Clustering hence can be regarded as problem of multi-dimension space, an extension of plane problems. Clustering can be applied on different disciplines. For financial companies, clustering help classify different borrowers in terms of credit risk level. The company can then offer different lending rates and borrowing ceiling to different borrowers according to their risk level in order to strike a balance between profit and risk. 一、聚類 (Clustering) 一 條 數 據 紀 錄 可 能 包 含 多 個 屬 性 (Attribute)。以一所財務公司的客戶資 料為例,其客戶數據紀錄包括客戶的年 齡、職業、薪金和貸款紀錄等資料。聚 類可按數據間的距離將其分類。如數據 只有兩個屬性,計算數據距離的方法便 像計算二維平面上兩點的距離。因此, 聚類可視為是多維空間的問題,為平面 問題之延伸。聚類的應用很廣,以財務 公司為例,聚類可把借貸人分類成不同 的信貸風險水平,以便訂立不同的借貸 利率和借貸上限,在利潤和風險之間取 得平衡。 16

(17)

Algorithms above are too simplified and cannot deal with every data set. Clustering, as an instance, if some data contain extreme values, i.e. very large or small value, cluster may deform and performance of model significantly decline. In order to cope with data with different distribution, new analytic algorithms are introduced in order to discover more useful information.

以上的算法頗為簡陋,不能應用於所有數據 組合。以聚類為例,若數據出現極端數據, 即含極大或極小數值的數據,進行聚類分析 時可能使各群組變形,大大削弱分析效能。 為了應對不同的數據分佈,不斷有新分析算 法推出,以便發掘更多有用的資訊。 二、最近 K 項鄰居法 (K-nearest Neighbor) 這 是 另 一 應 用 數 據 距 離 概 念 的 數 據 掘 挖 算 法,用於預測數據未知的目標屬性 (Target attribute)。此算法首先計算已知目標屬性數 據和被預測數據甲的距離,再抽出與甲最接近 的 K 個數據,按它們的目標屬性以少數服從多 數的方法,決定甲的目標屬性。因為 K 可以是 大於零的整數,所以需要測試有不同 K 值的模 型準確性,以決定最佳的 K 值。 3. Association Rule

Association rule is commonly used to establish cross-selling strategy. In retailing industry, there is a purchasing tendency between individual goods. For example, customers in supermarket who has bought bread may have high tendency to buy milk. Those goods can be put together to stimulate customers making more consumption. Association rules base on probability can identify these relationships. From this algorithm, we can obtain following three reference values:

Suppose we examine sale relationship of Good A induces Good B—

Support: probability of two goods being purchased at the same time;

Confidence: given that Good A is purchased, probability that Good B is also purchased;

Lift: "Confidence" of that relationship divided by the probability Good B is bought.

If three values are at a high level, the association rules can be used for setting cross-selling strategies.

三、關聯規則 (Association Rule) 關聯規則普遍應用於訂立交叉銷售 (Cross Sale) 的策略。在零售業中,個別貨品的銷 售當中存在傾向性。例如顧客一般在超市 購買麵包時,亦有很大傾向購買牛奶。只 要把這些配對的物件放在一起,便可刺激 顧客作更多消費。關聯規則可透過概率識 別這些關係。由此算法得出的參考數值有 三: 假設現在要探討甲貨品引發乙貨品銷售的 可能— 支持度 (Support),即甲乙兩貨品同時出現 於同一交易的機率; 確定性 (Confidence),即假定甲貨品被 購,乙貨品也被購的機會率; 增益 (Lift),「確定性」除以貨品乙被購的 機會率。 如三者數值都處於高水平,該關聯規則便 可用於設定交叉銷售的策略中。

One application of K-nearest Neighbor model is determining whether a

company will go bankrupt by using various financial ratios as input data. 最近 K 項鄰居法能以各項財務比率作輸入數據,預測某公司會否破產。

2. K-nearest Neighbor

This is another algorithm applying data distance concept, which is used to predict unknown target attribute of data. The algorithm first calculates distances between being forecast data A and all known target attribute data. Then the K nearest data neighbors will be extracted. According to their target attribute, by voting basis, target attribute of A will be determined. As K can be any positive integers, accuracy of models with different K values will be tested, in order to obtain the optimum K.

(18)

Well-recognized

Risk Management Certificates

Testimonies to Competency

Have

you ever thought of being a member of the world-wide elite group? Would you like to be recognized as a world-wide professional financial risk manager? If yes, we would like to introduce to you two risk management examinations:

FRM(Financial Risk Manager) and

PRM(Professional Risk Manager). We will focus on the basic information and comparisons on these two examinations.

Please refer to the following tables for comparison: FRM PRM GARP (Global Association of Risk Professionals) Website: http://www.garp.org/ frm/overview/overview.aspx PRMIA

(Professional Risk Managers' International Association) Website: http://prmia.org/ INSTITUTION EXAM FORMAT STRUCTURE EXPECTED PASSING RATE EXAM ORDER RESULT RELEASE DATE VENUE IN HK

Paper and pencil multiple choice

2 Exams Part 1 then Part 2 or both Part 1 + Part 2 at the

same time November and May

Asia World Expo, HK International Airport

(for 2011) 4 to 6 weeks after the exam

About 50% About 50% Computer-based

multiple choice 4 Exams No specific exam order

At almost any time throughout the year 10 test centres in HK

15 days after the exam

F o r F R M , y o u m u s t pass exam part I first before taking exam part II. NOTE: If you take BOTH part I and part II on the same date and fail Part I, Part II will not be graded. If you failed in Part I, you would have to register again for both parts of the exam. EXEMPTION None Exemption is possible. Please refer to Q2 in ‘More Information’ FRM PRM PRM FRM 18

(19)

Q1 Are there any discounts available?

FRM: No.

PRM: Discounts are available for groups of 10 or more candidates from the same organization.

Q2 Any exemption?

FRM: No exemption.

PRM: Use the table below to identify which cross-over exemptions you qualify for and which exams of the PRM program you will need to pass to attain the PRM designation.

Program Exemptions Register for These Exams

CFA, CIIA, CEFA, FSA, CQF Exams I and II Exams III and IV CSI(Financial Risk Management), CAIA Exam I Exams II, III and IV ASA Exam II Exams I, III and IV Associate PRM certificate Exam IV Exams I, II and III

You may combine cross-over credits. For example, if you are in ASA and have passed the CSI, you are exempt from both Exams I and II.

Q3 Can I bring a calculator or PDA into the exam room?

FRM: Only FRM authorized calculators are allowed.

The calculators are listed at http://www.garp.org/frm/overview/policies/calculator.aspx

PRM: No. An online scientific calculator is part of the test delivery system. No other materials could be brought into the exam room with you.

None PREREQUISITE None

DESIGNATION REQUIREMENTS

1. Be an active member of GARP 2. Pass FRM exams

3. Submit two-year related working experience within 5 years after you have passed the exam

1. Be an active member of PRMIA 2. Pass PRM exams

3. Commit to uphold the highest professional and ethical standards as defined by the PRMIA Standards of Best Practice, Conduct and Ethics

REGISTRATION Online or via fax Online

More Information

PRM PRM PRM FRM FRM FRM

All the information in this section is for your reference only as the examination requirements and arrangements are subject to change. For more accurate and updated information, please visit the organizers’ websites.

參考文獻

相關文件

{ Title: Using neural networks to forecast the systematic risk..

僱主 工作者 教育

– 某人因為與上市公司有關連,即內幕人士 (如公司董事、職員或公司的 核數師等)

– 第一百八十四條之一規定:雇主使勞工使用 危險物從事作業前,應確認所使用物質之危

圖4 1 整合資訊系統風險 圖4.1 整合資訊系統風險..

李佳芸 保險金融管理系 商業與管理群 已報到 陳昶霈 保險金融管理系 商業與管理群 已報到 陳錫建 保險金融管理系 商業與管理群 已報到 潘學恩 保險金融管理系

林旻柔 保險金融管理系 商業與管理群 已完成網路報到,且收到考生畢業證書 王美晴 保險金融管理系 商業與管理群 已完成網路報到,且收到考生畢業證書

 包括 包括 包括 包括學校 學校 學校管理 學校 管理 管理、 管理 、 、學生學習 、 學生學習 學生學習、 學生學習 、 、校風 、 校風 校風及其 校風 及其 及其 及其 他支援. 他支援