RMBI NE SlETTER
RISK MANAGMENT AND BUSINESS INTELLl GENCE, HKUST
DECEMBER. 2011
Director's Column
Time to put into practice what has been learnt
Time flies. BSc in Risk Management and Business Intelligence (Rl\個1program), the frrst undergraduate program in Hong Kong to integrate risk management and business intelligence, has been running for over two years. Our frrst batch of elites has entered their final year of study in September.
To equip students with solid knowledge in risk management and business intelligence, we emphasize both hard skills and soft skil1s. For hard skills, we make good use of interdisciplinary resources to strengthen students' analytical ability by sharpening their proficiency in statistics, mathematics, fmance, and information systems. Soft skills are equal1y important. Therefore, we organize various enrichment programs in the forms of experience sharing sessions, workshops, seminars, and conferences for our students to leam 企omexperienced industry practitioners and look into business problems 企ommultiple perspectives.
Time is ripe for our first batch of elites to 甘yth巳irhands at real business problems in risk management and/or business intelligence, and to demonstrate their capability in problem solving by utilizing what they have leamt from the RMBI program. Nine capstone projects covering various risk management or business intel1igence topics in the fmance, retail, healthcare, and information technology industries have been carried out in this academic year. With the guidance of our university professors and industry professionals, 1 am sure that our students will not only gain valuable experience in handling practical business issues in risk management and business intel1igence, but also be able to provide useful recommendations to project providers.
In this issue of our newsletter, we invited senior staff 企oma prernier theme park and a major public utility company in Hong Kong to share their vi巳wson developing an enterprise risk management system. You will also find int巳resting artic1es and hot topics in risk managem巳ntand business intelligence in the newsle前er.1 hope you enjoy reading it.
適時學以致用 時光飛逝，作為全港第一個結合風險管理和商業智能的學士課程， 風險管理及商業智能學開辦至今已逾兩年。我們的第一批精英巳於 九月踏入他們課程中最後一個學年。 為了讓學生在風險管理與商業智能方面擁有扎實的知識，我們對硬 性和軟性技能同樣看重。硬性技能方面，課程善用跨學院的資源，強 化學生在統計、數學、金融與資訊科技方面的能力。軟性技能同樣重 要，所以我們舉行了大量增潤活動，包括經驗分享、工作坊、請座和 會議，讓學生從經驗豐富的業界人士身上學習，並由不同角度創析 商業問題。 現在時機已成熟，是時候讓我們第一批精英實踐在課程中所學到的 知識，協助解決風險管理和/或商業智能範i 尸 疇的商業問題，展現他們的解難能力。九個 涵蓋金融、零售、護理與資訊科技範疇的風 險管理和商業智能專題研究報告巳於本學1 年進行。在各大學教授和公司的專業人員的 指引下，我確信學生不僅能吸取處理真實風a 險管理及商業智能商業問題的寶貴經驗，也 可向項目提供者給予有用的建議。 今期新聞報會有香港著名的主題公園和公 用事業公司分享他們發展企業風險管理系 統的觀點。你亦可找到最新而有趣的風險管 理及商業智能主題文章。希望你會喜歡本期 的新聞報!
Contents 目錄
Director's Column HKUST RMBI NewsleUer Editorial Board
Feature Story
3
Advisors Enterprise Risk Management in Theme Park
Ocean Park Hong Kong
Prof Mike K.P. SO
主題樂園的企業風險管理-香港海洋公園
(Program Director of RMBI) Professional Column
7
Dr. Adela S.M. LAU Introduction of Public Utility Risk Management (Senior Instructor of RMBI) TowngasMiss Sandra M.K. WONG
公共企業風險管理簡介-煤氣公司
(Executive Officer of RMBI)
Interview
11
Editors: BI in supermarket industry An Interview with the Data
Analyst in a Large Retail Chain Store
CHAN Hiu Tung
,
RMBI year 3 大型零售店數據分析師專訪商業智能在超市業的應用CHAU
Li
bby Lok Yee,
RMBI year 2 Risk Management Digest14
CHUN Pui Ying
,
RMBI year2
LAM Hing Ki,
RMBI year 3 Algorithmic Trading NG Wai Ho,
RMBI year2
算法交易 PENG Ming Chuen
,
RMBI year 3YIM King Pong
,
RMBI year2
Business Intelligence Digest17
Introduction to a common Business tool (PALO)
通用商業工具介紹 Reporters:
21
Glossary KONG Chun F剖， RMBI year
2
LAU Ka Ming,
RMBI year2
22
TAM Hiu FungNG Man Hin,
RMBI year,
RMBI year2
2
Program DevelopmentPublication December
,
2011
Contact
Risk Management Business Intelligence Program The Hong Kong Universily of Science and Technology
香港科技大學風險管理及商業智能學課程
Phone: 2358 6955 Fax: 2335 9317 Email: [email protected]
Website: http://www.rmbi.ust.hk
。2011 RMBI
,
HKUST.AII
rights reserved.Feature Story
Enterprise Risk
Management in
Theme Park
:
Ocean Park Hong Kong
In 2003, the outbreak of severe
acute 附p廿atory
syndrome (SARS) dealt a severe blow to tourism in Hong Kong,
and the theme park industry was especially hard hit. Fortunately,
riding on the economic benefits brought by the mainland tourists,
Ocean Park Hong Kong,
a marine life theme park,
overcame the difficulties and challenges posed by SARS and was named by Forbes as the World's Seventh Most Popular Amusement Park. In 2004,
the Park launched a HK$5.5 billion development and construction project,
the largest ever in its history. In order to ensure a smooth implementation of the project and an e仟'ective risk control,
the Park has set up an Enterprise Risk Management System (ERM system), which is a relatively innovative concept to theme park industry and its staff as well as in Hong Kong. We are glad to have invited Mr. Frankie Hau, Senior Manager of Environment, Health and Safety and Miss Vivian Lee,
Marketing Director of the Park,
to give us an introduction to the system and the risk management measures in respect of operation,
safety and business sales.Importance of Enterprise Risk Management
Risks can expose a company to unanticipated consequences like profit and investment losses. Enterprise risk management is a measure to manage the company by identifying, analyzing and mitigating risks. Mr. Hau emphasized that enterprise risk management covers not only management of risks related to health and safety issues
,
but also takes into account areas like financial risks,
operational risks and pr吋 ect risks. A systematic and efficient enterprise risk management raises the performance of employees and the quality of the Park's operation,
and also prevents and reduces the losses incurred when accidents happen.The Park has experienced various difficulties and challenges during its 30 years of history in Hong Kong. The management has therefore accumulated a lot of valuable experience. "The systematic integration of past experience and consolidation of the existing control measures have greatly enhanced the management quality and achieved better planning on resources allocation. And that is exactly the concept of implementing an ERM system in Ocean Park." Mr. Hau said.
Mr. Frankie Hau, Senior Manager of Enviroment, Health and Safety and Miss Vivian Lee, Marketing Director of Ocean Park
海洋公園璟境健康及安全高級總理侯文強先生和市務總監季玲鳳小姐
The first theme park to establish an ERM
system
Ocean Park is the first among its peers to establish an ERM System
,
and consultants were employed from Australia to assist in the coordination and roll-out ofthe system. With the help of the ERM system,
Ocean Park is able to perform a variety of tasks like categorizing and establishing control measures,
conducting regular reviews on the risks profiles and the control monitoring, as a result, submitting the annual board report.The ERM system operates by identifying different risk event and classified them into separate categories. Corresponding control measures will then be drawn up and staff will be designated to take charge of the follow-up work. "For those events that Ocean Park has not experienced before
,
such as tsunamis,
earthquakesFeature Story
and other natural disasters and risks that may arise from development and investment in new projects
,
Ocean Park will deal with them upon identifying foreseeable risks events. With the assistance of the ERM system, the Park can develop a series of control measures specific to each of the risk event. If unfortunately accidents should happen, the Park can respond promptly by carrying out the relevant contingency measures in order to minimize the impact." Mr. Hau explained.Besides
,
the Park has set up a Risk Management Committee comprising the Chief Executive Officer,
Chief Financial Officer and all executive directors. The committee meets every half year to update on the implementation of the ERM system and to review any changes in the risk profile and any major events happened in the Park during the half year,
so as to develop policies for improving the existing system. AII decisions made by the committee would be disseminated to the frontline staff to ensure the system is being executed e仟'ectively.Examples of the ERM system functioning in areas of operation, safety and business risk management in Ocean Park are as follows:
Operational Risks
Visitors flow:
A low visitor count will a仟:ectthe profit of the Park
,
whereas too many visitors will overload the facilities and may pose safety risks. Through the ERM system,Ocean Park controls visitors flow by adopting an appropriate approach in the form of crowd control measures,
emergency response plans and queue control, depending on the circumstances at the time. Take queue control as an example,
it varies in different seasons or festivals because special events will a仟'ect the distribution of visitors in the Park. During special events such as Summer Splash 2011,
it was expected that Rapids Adventure and Raging River would attract hordes of visitors while leaving traditional facilities such as Merry-go-round with a low traffic,
and therefore special arrangement would be needed to maintain proper queue management. Ocean Park also works closely with Citybus, the operator of bus route from Admiralty MTR Station to Ocean Park, to maintain proper crowd control. By sharing data on the number of visitors,
Citybus is able to arrange an appropriate timetable of bus service and transport visitors to the Park orderly. Meanwhile,
Ocean Park will set up suitable arrangements at the entrance and install signposts in the Park to direct the visitors to different places.Animal care:
The wide variety of animals in Ocean Park gives it a special appeal as a popular tourist spot in Hong Kong. Yet, the deaths of Chinese Sturgeons in the aquarium have raised public concern over animal care in Ocean Park.According to Mr. Hau
,
Ocean Park has always been committed to providing the best care to its resident species. The Park would also seek help from experts when new species are introduced to its animal collection in order to give proper care for the. new members. Apart from regular healthcheck-u 阱， the Park also emphasizes on animal care practices such as operations and trainings on animals. AII these control measures have been included in the ERM system as general guidelines.
Safety Risks
Maintenance:
Maintenance is another critical factors contributing to the success of Ocean Park. Ocean Park has regular and yearly maintenance policy in place to ensure high standard of safety of its facilities. A comprehensive response plan is also incorporated into the ERM system to handle major incidents such as the collapse of facilities or power failure. For instance
,
Cable Car will not be allowed to operate when wind speed exceed a certain limit. And during the temporary shutdown of Cable Car, corresponding measures as to how the announcement is to be made and how to direct visitors to take other means of transportation will also be carried out. Should there be any accident,
like the one on the Ocean Express last year,
the Park will conduct a series of reviews on the safety of the facilities and the staff practices to find out the cause of the accident. Moreover,
measures will be set up via the ERM system in order to prevent similar incident from happening again.Business risks
Sales and Marketing
The greatest risks to the operation comes from business risks
,
which are mostly related to the sales and marketing division. Typically, the biggest threat in this division is the lack of visitors,
which usually happens when disaster strikes, such as SARS in 2003. The ERM system maintains a set of measures addressing those risk events, providing discounts and having joint promotions in off-season periods are among the most common practices. More and more theme parks are being built in Asia and thisFea
tur
e S
t
o
ry
may threaten the status of Ocean Park
,
as some may think. In response to this,
Miss Lee pointed out that Ocean Park has a different focus of positioning which emphasizes on the connection between human and nature,
and it aims to provide visitors with a happy,
unforgettable and meaningful experience that combines entertainment and education. Besides,
with the launch of new facilities after completion of the expansion project,
visitors will have a greater sense of the uniqueness of Ocean Park that makes it stand out among its peers.Difficulties,
OpportUriities,
Challenges
Installation of a new system in a company often means facing up to different kinds of difficulties. Mr. Hau said there had been a shortage of experienced frontline staff during the early stage of the implementation of the ERM system. However
,
Ocean Park did not face many challenges as the systerri was executed with commitment from the top management. Also,
even though some employees may not understand the whole structure of the ERM System,
they can still perform well in their duties and assist in the smooth operation by giving them clear instructions and maintaining a close communication with them.Having been in operation for nearly five years
,
the ERM system has not only enhanced operational efficiency in the Park but also promoted information exchange across different departments,
which are positive to risk control and management. With the number of visitors expected to rise after expansion of the Park and industry competition intensifies,
the ERM system will play an even more substantialrole in the Park's operation.To overcome new challenges
,
every successful company needs to keep abreast with the time in all aspects. In addition to being one of the most popular theme parks in the world, Ocean Park has moved to become a pioneer in the development of the ERM system. Leveraging on advanced technologies,
the Park has succeeded in mitigating risks in operation,
business reputation and marketing, while enhancing service quality and smoothing operation processes. This innovative concept of enterprise risk management by Ocean Park can serve as a useful guide for lifting the standard of amusement parks around the world.主題樂園的企業風險管理一
香港海洋公園
2
。阱，非典型帥炎(沙士)大規模爆發嚴重打擊了香港的
旅遊棠，主題公園這個行業更是飽受衝擊.幸而受惠於內地迦客 所帶來的經濟效益，以海洋生物為主題的香港海洋公園克服了沙 士所帶來的種種困難和挑戰，更無謹福布斯網站評選為「全球第 七大最受歡迎的主題公園 J ﹒ 2004年，海洋公園開展其有史以 來最龐大的發展及擴建計創，投資額高違 55億港元.與此同 時，為確保計劃能順利進行及有效控制園區內各項風險，海洋公 園設立了企業風險管理系統 (Enterprise Rlsk Management System ，簡稱 rERM 系統 J ) .這個系統對海洋公園的員 工、主題公園業界及香港而奮，都是頗為創新的概念.我們很榮 幸能夠邀請到海洋公園環攪健康及安全高級經理侯文強先生和市 務總監李玲風小姐為我們介紹該系統和公園在運作、安全和業務 營銷方面的風險管理措施. 企業風險管理的重要性 風險可以為公司帶來難以預料的後果，如盈利虧損、投資失 制，因此有效控制風險對一間公司來蛻是十分重要.企業風險管 理便是一個以融別、分析和降恨風險為基礎來管理公司的措 施.候先生強調，企業風險管理不懂涉及健康和安全風險管 理，也包括其他範團如財務、營運和項目風險管理等.有系統和 高效率的企業風險管理不但有助提升員工的工作表現及園區的營 運質景，還可及早預防和減恨發生事故時所造成的損害，對企業 的畏連發展有莫大轉益. 海洋公園在香港已有三十多年歷史，過去曾經歷過種種困難和挑 戰.因此，管理層累積了很多寶貴的管理經驗.侯先生表 示 r 將過去的經驗有系統地整合後，用於鞏固現有的控制措 施，可以大大提高管理質索，並使資源分配的規劃更為妥善，這 就是海洋公園建立即M系統的理念. J 首個設立 ERM 系統的主題公園 海洋公園是主題公園業界中第一間建立即M 系統的公司，並特別 從澳洲聘請顧問來港協助籌劃及推行該系統.海洋公園借助該系 統來執行不同的風險管理工作，包括將風險分類、制訂控制措 施、定期轍討風險評級及提交董事局報告等.Feature Story
在運作土，系統會將訂定的風險事件分類，並制訂相應的控制 描施，安排專責人員跟進。侯先生稱 r 對於海洋公園未經歷 過的事故如海嘯、地價等自然災禍，以及在發展和投資斬項目 的過程中可能遇到的風險，海洋公園在職別出可預見的風險事 件後，便會著手處理.透過ERM系統的協助，海洋公園可針對 風險事件逐一制訂一系列調控措施.之後若真的不幸發生事 故，海洋公園便可迅速執行相闊的應變措施，就iLt事故所帶來 的衝擊 o J 此外，海洋公園已設立一個自行政總裁、財務總監和全體執行 董事組成的風險管理委員會，並每半年舉行一次會攏，目的是 擴晏員會了解 ERM 系統在公園的運作，同時檢討過去半年風 險狀況的種化和園區所發生的大型事件，以制訂政黨來改善 現有機制，而所有決黨也會下違前線員工，使系統能有敢地運 行. 以下是ERM 系統於海洋公園的營運、安全及商業風險管理方 面發揮成效的一些例子: 營運風險人流
遊客太少會影響公園的盈利，但遊客太多也會影響遊客車用各 頂設施，並構成安全風險.透過ERM 系統，海洋公園會棍乎 園區的擠迫情況探取適當的處理古法，如控制人流、實施應急 計劃和隊列控制策略.以隊列控制為例，在不同的季節或特別 節日，園區的遊客分布亦有不同.例如在夏水禮2011 這個活 動中，系統預計息流和滑浪飛船等水上設施會吸引大量人 流，而傳統遊樂設施如旋轉木馬的遊客則可能會較少，隊列管 理便要為此作出特別安排.在人潮控制方面，現時城巴設有來 往金鐘港鐵站至海洋公園的巴士線，海洋公園會眼城巴緊密聯 繫以便雙方掌握迦客數目的資料，之後蛾巴便可安排適當的巴 士班次，將遊客有秩序地帶到園區，而海洋公園則可在入口處 作出適當的入場安排，圍內也會設置各種指示牌，疏導遊客到 不同景點. 動物鐘理 香港海洋公園飼養了各 it各樣的動物，令這地方成為香港一個 別具吸引力的旅遊景點.然而，中華蟬的死亡事件曾經引起公 恩關注海洋公園在動物專業體理方面的能力.侯先生解釋，海 洋公園一直認真照顧各類動物，在引人新動物時也會與相關的 專家研究如何照料，而 ERM系統對各種動物的種理亦有清晰的 指引。除7 為動物進行定期的健康檢查外，海洋公園也十分注 重種理人員對動物種理的專業知攏，例如動物訓練和為動物施 于衛的知纖. 設施保饗 海洋公園的成功，設施保養得宜是另一項關鍵因素 o ERM 系統內 載有關於定期和年度維修的全面策略，確保園區能維持高水準的 安全標準.系統亦有應付設施倒塌或停電等嚴重事故的錯略.例 如當風速超過一定的限制，纜車便不可運行，以確保遊害的安 全.暫停纜車服務時亦會有相應措施配合，包括如何向旅客發佈 消息及工作人員如何指示旅客乘搭其他交通工具.若不幸有意外 發生，如去年涉及海洋列車的事故，公園會對設施的安全和員工 培訓進行一連串的調查，以找出導致該事故發生的原因，並會透 過 ERM系統制訂相關策略，避免同類事件再次發生. 商業風險 銷售及營銷 商業風險是企業營運的最大風險，道類風險大多涉及銷售和營銷 部門。入揚人次不足對海洋公園的營運威脅最大，道種情況通常 在社會發生大型事件時出現，如零三年的非典型肺炎事件 o ERM 系統備有一系列應付此等情況的措施，例如在淡季期間提供折扣 和聯合促銷優惠. E洲的主題公園不斷增加，有人認為這可能會對海洋公園構成威 脅.李小組回應指，海洋公圍著眼點與其他公園不同，海洋公園 強調人與自然的珊，巷，並將娛樂和教育連繫在一起，帶給迦客一 個開心、難忘和具教育意義的經歷.而海洋公園擴建後，將會有 更多的新設施，遊客將更能體會海洋公園與其他主題樂園的分別 和獨特之處. 困難、機遇、挑戰 一間公司在探用一個新系統時可能會遇上各種困難.候先生表 示，海洋公園於實施ERM 系統初期，缺乏熟悉系統操作的前線員 工.儘管如此，由於高層管理人員都十分支持推行該系統，故此 海洋公園在執行該系統時並沒有面對很大的挑戰.此外，雖然部 份前線員工不一定完全了解企業風險管理策略的整體結構，但管 理厲透過發出清晰的指示和與前線員工保持緊密溝通，轍員工清 楚知道自已的工作崗位，協助公園順利運作。 ERM系統已建立 7 近五年，不但提升了公園的營運效率，亦鼓勵 和加強 7 各部門的資訊交流，對海洋公園的風險控制和管理極為 有利.預期海洋公園擴建後遊客數目將會增加，加上業界競爭越 趨劇烈， ERM 系統在公園運作土將搪當更重要的角色. 每一間成功克服斬挑戰的公司，必需在各 13 面都能與時並進.世 界最受歡迎主題公園之一香港海洋公園開創業界先河，率先建立 本身的 ERM系統來管理風險.通過先進的技衛，海洋公園不但降 恨了營運、商業信譽和市場營銷的風險，同時更提高了服務質量 和使運作更為順暢。香港海洋公園這個創新的概念可作為世界各 地主題公園的借鏡，以提升全球主題樂園的整體水平.Profess
i
ona
l
Co
lumn
Introduction of Public Utility Risk Management - Towngas
Mr. Victor C L Kwong, Corporate Hea1th, Safely and Environment Manager
企業安全及風險管理經理郎超靈先生
In March 2011
,
northeast Japan was struck by a magnitude 9.0 earthquake,
the most powerful one ever recorded in Japanese history. The earthquake and the devastating tsunami that followed triggered a crisis at a nuclear power plant, threatening the lives of billions of people. Yet,
apart from the natural disaster,
inadequacy in risk management was partly to blame for causing the nuclear crisis. The disaster sparked public concern over risk management in energy safety-related issues. It is therefore necessary for us to outline the risk management system of the energy sector in Hong Kong so as to address the public concern.5afety and reliability are the top priorities for a public utility enterprise. To that end
,
it is imperative for us to maintain proper management on operational risks; otherwise public confidence will be lost. In general, the enterprise risk management system we adopted is based on 5tandard 150 31000,
the general guidelines on risk management promulgated in 2009. Before 2009, A5/NZ5 4360:2004,
the former version of 5tandard 15031000,
was in use as the blueprint of Towngas' risk management system.In gas industry, a common operational risk is third party damage of underground pipelines. Our gas production plants are built in 的xed locations. This has made plant management easier and more controllable despite the fact that large quantities of inflammable gases and liquids are stored at our plants. Our pipeline network is more vulnerable to third party damage due to its large coverage area in Hong Kong. In 2004
,
a gaspipe in the capital of Belgium exploded as a result of third party damage
,
causing over a hundred of deaths and injuries. Hence,
safety management of pipelines is a major concern and we havé taken various measures to mitigate risks in this regard. For example:1. Our safety steering committee reviews and ranks the risks regularly. Each risk is monitored and managed by a designated staff
,
to ensure that all risks are under control. 2. To cultivate a safety mindset from the top down,
regular safety talks are given by the senior executives to the general staff with an aim to avoiding incidents similar to the oil spill in the Gulf of Mexico in which hasty examination of the oil gusher failed to prevent the accident from happening.3. We are also stepping up efforts in educating external organizations on pipeline safety.
4. We are also conducting trench inspections to avoid pipelil:le damage due to road works. Besides
,
regular preventative inspections of pipelines are being éondusted to identify aging pipelines,
if any at an early stage.As a result of our efforts
,
the number of accidents caused by third party damage dropped significantly to 11 in 2009 from more than 120 in 1997.To maintain the reliability of town gas suppl比 weoperate two gas production plants in Hong Kong
,
one in Tai PO and the other in Tokwawan. Town gas is mainly produced byProfessiona
l
Column
the Tai PO plant. The Tokwawan plant is mainly operated in peak seasons and hours to cope with any sudden increase in demand.ln addition
,
to ensure stability of our feedstock supply, we import natural gas from Guangdong Dapeng
LNG Company Ltd in Shenzhen as a feedstock for gas production while continuing to produce town gas from naphtha at our Hong Kong plants. This dual feedstock strategy prevents shortage of feedstock as a result of accident in the Shenzhen natural gas plant.Towngas has been actively seeking ways to ensure the reliability of feedstock supply for gas production with a view to providing economic bene1its to our customers. Since Oct 2006, we start producing town gas from dual feedstock, i.e. naphtha & natural gas. We made a 25-year natural gas supply contract agreement with an Australian natural gas company in 2003. The price we pay for natural gas is approximately linked to crude oil price of USD$25 per barrel. It is de1initely a good price to us with reference to the market value of crude oil nowadays. It stabilizes our gas production cost
,
and further ensures the reliability of our feedstock. Thus, an operational risk is reduced. As well,
this enables us to lower the tari仟 toour customers by more than 10%.We have more than 90 gas pipeline pr吋 ects in mainland China. A safety and risk management scoring system has been set up with the purpose of rating each joint venture and assessing the major risks for each of them. Safety is given the heaviest weighting in the overall assessment. In December of 2003
,
a gas leakage incident in a mine in Chongqing operated by PetroChina Company Li mited claimed the lives of more than 230 people. The incident once gain illustrated the importance of supervision & non-standard-operation. The scoring system is a useful tool to assess if a joint venture company meets our risk requirements. We also organize business risk assessment workshops regionally to educate the joint ventures on the latest trends in risk management and provide practical examples for them to improve their operations.Before making an acquisition in mainland China, we must send our staff to conduct site inspections and examine their documents in order to get a clear picture of the potential risks involved. In our mainland joint ventures, we have our own direct staff for the key management positions
,
i.e. General Manager, Engineering Manger
&Financial Manager to ensure that our policies
,
work procedures & standards are properly followed. Otherwise,
operational risks are difficult to be monitored and controlled. Thus,
investment risks are high. Our management rights in joint venture are critical and it helps reduce the 1inancial & investment risks in mainland China.To our customers
,
safety,
reliability of town gas supply and the quality of customer service are the main concerns. Our town gas supply achieves a near 99.99% reliability and the possibility of supply disruption is extremely low. If unfortunately an accident,
such as damage to the pipelines,
occurs and leads to temporary breakdown of town gas supply,
our target is to restore gas supply within 12 hours. Repairing and maintenance services are provided to our customers. Our emergency team could arrive on site within 30mins (> 90% request),
upon emergency enquiries concerning gas safety. We are com-mitted to satisfying customers' needs and have put in place various measures to minimize customer-related safety risks.Professional Column
As a leading public utility
,
Towngas has launched a series of programs to ful刊 1 our social responsibility to protect the environment.1. In 2003
,
Towngas launched the "Gas Appliance Recycling program" to collect old appliances disposed by customers and recycle the scrap metal retrieved. AII proceeds from the program are allocated to a green fund to support environmental activities in Hong Kong. 2. In 2006,
Towngas built a plant near a landfill in Tak Ku Li ng to collect methane from the landfill as burning fuel for gas production.3. In Oct 2006
,
we have started using natural gas as one of feedstock. It further reduces the consumption of naphtha as well as emission of greenhouse gases.4. In addition to reducing emission
,
we are keen to promoting recycling in the community and in our daily operations.Over the past five years
,
we have reduced overall carbon dioxide emission by 20%. We are now preparing a carbon footprint index on customer number basis for reference. To conclude, we face different types of ri時， includingoperational risks，的 nancial risks & environmental risks in our daily operation. We need to adopt a variety of mea-sures to hedge against those risks.lt is impossible to elimi-nate all risks but we strive to control and minimize them in order to provide our customers with safe & reliable supply of energy and caring services
,
while working to preserve our environment.The monitor shows number 01 calls phoned in and number 01 staffs on duty.
螢光幕顯示已接聽電話的數目沒當值人員的數目
公共企業風險管理筒介一煤氣公司
螺氣
Towngas
二零一一年三月，日本東北部遭該國有記錄以來最強的九 級地震和毀滅性的海嘯所衝擊。地震引發的核電廠危機威 脅數以十億計人口。然而，除 7 自然災害，風險管理工作 不足亦是構成核事故的部分原因。 這場災難引起公眾對能 源安全風險管理的關注和接慮。因此，我們有必要說明香 港能源業的風險管理制度，以釋除公眾的憂慮。 安全及可靠是公用事業機構的首要目標。因為缺乏其 一，都會令市民對企業失去信心。我們必須要善管理營運 風險，以達致這兩個目標。整體來說，我們的企業風險管 理系統都是依照二零零九年的 IS031000標準建立，這是一 套風險管理的通用準則。二零零九年前，煤氣公司是用 IS031000標準的前身AS/NZS 4360:2004作為藍本，制定 我們的風險管理制度。 燃氣行業最常見的營運風險來自地下管道。煤氣公司的廠 房雖然儲存了大量易燃氣體和液體，但由於廠房建於固定 位置， 因此相對地較易控制及管理。地下煤氣管道覆蓋全 港，容易遭第三者破壞。在二零零四年，位於比利時首都 的輸氣管便是因為第三者破壞而發生爆炸，造成過百人死 傷。因此，管道安全管理是煤氣公司最關心的環節，我們 亦採取 7 很多措施減f~此方面的風險。如: 1 .設立安全督導委員會會定期檢視各種風險，為風險評 級。每項風險由專人負責監察及管理，以防意外發生。 2. 高層會定期向員工講解安全的重要性，實行上行下效，避 免像墨西哥灣漏油事故中，高層到鑽油台草率檢查，錯失 阻止意外發生的時機。 3 向其他機構加強煤氣管道安全教育及檢查，以兔其他地下 工程意外損壞輸氣管而發生意外。 4. 定期主動檢測輸氣管道，展早發現管道老化，防止氣體洩 漏。 在煤氣公司的努力下，地下管道受第三者破壞等意外的數 目，成功由一九九七年的一百二十多宗，大幅減至二零零 九年的十一宗。Professional Co
l
umn
至於維持煤氣供應穩定方面，蝶氣公司在港有兩座煤氣生產 廠房，分別位於大埔及土瓜灣。現時，大部分時間都是由大 埔廠房供氣，土瓜灣廠房在用氣高峰期才敵動運作，以避免 因需求突然增加而導致供氣不種定。男外，在獲取穩定的原 料供應芳面，煤氣公司一方面從位於深圳的廣東大鵬天然氣 公司天然氣廠房輸入天然氣以生產煤氣，另一方面繼續於香 港的兩座廠房利用石腦油生產煤氣，以防止深圳的廠房因發 生事故而導致原料短缺。 煤氣公司 一向積極尋求經濟的能源供應，確保供氣可靠 性，以及為客戶帶來經濟效益。公司於2006年 10 月，開始採 用天然氣和石腦油雙原料生產煤氣。然而，我們早於2003年 與澳洲天然氣供應商簽訂 7 長達 25年的供氣合約。在簽訂供 氣合約時，定價是以當時每桶石油 25 美元為參考，遠f底於現 時石油價格。這有助穩定和減 f~ 煤氣公司現峙的製氣成 本，大大降f~7 營運風險。我們透過幣料調整費機制，將減 省回饋客戶，客戶的煤氣費有超過 10% 的節省。 至今，我們於內地的管道燃究項目違九十多個。公司特別建 立 7一個安全及風險管理的評分制度，為每間合資公司評 分，以評估該公司的主要風險，而安全是評分的最主要項 目。於二零零三年十二月，中國石油天然氣股份有限公司於 重慶開縣鑽探礦井時，因缺乏監督，未有依照標準程序工 作，導致氣體洩漏，造成超過二百三十人死亡。由此可 見，妥善的監督對於投資內地公司十分重要，評分制度正正 司評估該合資公司是否符合我們的風險要求。此外， 我們會 在每個分區舉辦商業風險自評工作坊，說明風險管理的最新 情況，並提供一些實例，幫助他們改善運作。8taffs are listening the customers' calls (hotline).
當值人員正接聽顧客的查詢電話。 當我們收購一間內地公司，定必會派遣員工到實地視察，調 查帳目以 7 解潛在的風險。內地的合資公司大部分都是自我 們調派高層管理人員，如總經理、工程總監和財務總監等到 當地負責管理，以確保該公司跟從我們的管理政策、制 度、模式和規範等各芳面。螺氣公司認為，若果我們沒有管 理權，合資公司的營運風險難以監管和保証，這會令煤氣公 司的投資風險亦大大增加。因此，合資公司管理權十分重 要，並有助減f~我們在國內投資的財務風險。 客戶方面最關注的，就是煤氣供應的安全、可靠性和客戶服 務質素。媒氣公司的煤氣供應司靠程度為百分之九十九點九 丸，煤氣供應突然中斷的可能性極慨。假如不幸有意外發 生，例如喘氣管道遭破壞以致煤氣供應暫停，煤氣公司的目 標是於受破壞之後+二小時內恢復喘氣供應。至於顧客服 務，煤氣公司為客戶提供修補服務和保養工作。有關於燃氣 安全的查詢，煤氣公司維修隊可在客戶來電後三十分鐘 內(超過百分之九十的個案)到訪作檢查和維修。煤氣公司重 視客戶需求，並訂定措施以降 f~與客戶相關的風險。 作為一間公用事業機構，煤氣公司展開 7 多項計劃以肩負其 保護環境的社會責任。包括: 1. 二零零三年，煤氣公司推行「點廢成金舊爐具回收計劃 J '這計劃會收集客戶的舊爐具並回故當中有用的金屬，所得 款項會全數撥入煤氣公司的環保基金，用以支持香港環保活 動。 2. 二零零六年，煤氣公司在新界打鼓嶺的垃圾堆填區附近興 建工廠，收集堆填區內產生的沼氣作做料。 3. 二零零六年十月，煤氣公司開始使用天然氣生產煤氣，大 大節省石腦油用量，從而減少溫室氣體及污染物的排放。 4. 除 7 減少廢氣排放，煤氣公司亦推動社會和公司內部的廢 物回收工作。 煤氣公司過去五年的整體二氧化碳排放量大大減少了百分之 二十，而且現正準備提供客戶碳排放量指數作為參考。 總括而吉，煤氣公司一直面對許多不同類型的風險，包括營 運、財務和環境風險。我們必需採取多項措施來應付這些風 險。要完全消除所有風險是不可能的事，但我們會盡量減慨 和控制風險，為客戶提供安全可靠的能源供應和貼心的服 務，同時保護環境。
Interview
81 in supermarket industry
An interview with the data analyst in a large retail chain store
Q: How much resource do supermarkets put int。 the analyzing department?
A: Although resources being put into the analyzing depart-ment in supermarkets are less than that in the banking
secto巳 thisdepartment has gained importance in the past few yearsinthe supermarket industry with the wealth of data collected.
Q: How do supermarkets collect data for analysis?
A: There is a lot of ways to collect data, such as reward points program of loyalty card
,
Point of sale (POS) or checkout, e-commerce where supermarkets can track the sites customers have browsed and observe the products they are interested in, etc. Supermarkets in other countries have also experimented with analyzing CCTV footage tostudy customers' behaviour
, though
this is not popular in Hong Kong. Among these methods, POS is the most common practice as data collected at checkout is readily available, and it directly reflects the sales and profits. Infor-mation from reward points programs of loyalty 臼 rd will also be very useful for future use. AII the data collected helps us understand customers' buying behavior.Q: Can data collected be categorized for analysis?
A: Supermarkets collect different kinds of data every day.
There are no standards for categorizing data. Some types of data are for specialized reports, such as time series and
categorical data. A regular (this week vs. last week vs. last year/monthly/rolling-12 months) reports are generated by the computer system with the data collected to judge the soundness ofthe business.
Q: What are some of the indicators supermarkets lookat?
A: Different departments in a supermarket look for different indicators. For example, trading department looks at sales
per day and sales per linear footage, merchandising and store planning will consider sales per square feet as their prime indicators, supply chain may look at stock on hand data. Thus, there are many different types of indicators to
reflect the health of business.
Q:
Besides those regular inputs of data for the report,
how do supermarkets investigate other data? Are there anyexamples and what other data may be involved?A: Supermarkets may assign a team or outsource a
pr吋ect from time to time
,
to investigate other data besides those input regularly for the report. This is because our in-house developed data mining skills are still not in-depth enough and outsourcing the data toexperts can help us dig out more findings.ln recent years, there is more data from internet trading and e-commerce. New channels such as QR code and com-bining traditional print marketing also help explore
customers' needs in depths. This trend is expected to continue with outside agency.
Q: What are the major difficulties that a supermarket may face while implementing data mining in the operation?
A: It would be hard to align with the company's goals, this
in itself is already a very tough mission, if not impossible. A supermarket has numerous legacy systems and applications
,
how to synchronize them to make use of the data efficiently is a big problem.We can easily collect vast amount of data in our daily operations from points of sale, customer loyalty program, the log of our web server or even by observation. However, all of these are collected for promotional propose, or simply for calculating some Key Performance Indicators (KPI) as mentioned, which are merely at a surface level. The importance and usefulness of Business Intelligence (BI) were not well recognized and received,
or at least not until recent years. Thus
,
it would be very difficult for supermarkets to adopt datamining using the existing systems and applications.If we want to develop on the BI aspect, we would need to redesign the database structure, schema, and the Data-base Management System (DBMS). It would be a near-complete re-engineering of our operation system. The resources needed would be undoubtedly huge. We also have to risk bearing a loss-of-sales-cost if things go 0仟
track because the process will surely impact our daily operation.
Interview
Developing on the 81 aspect means we are digging deeper into customers' data. To accomplish this
,
we can either outsource the pr吋ector make use of the data by ourselves. We will need to be very careful with the privacy issue when making business decisions related to the findings concluded from our data.Another inconvenient truth is that
,
even if we pour resources into the data mining proce白，we might not be able to get the expected return sometime.s. For example,
say we use 81 to 的 nd out two items' relation using the Association Rule,
like the famous example 8eer and 8aby's diaper; even if we find out their correlation is high,
it does not necessarily mean that we can bond them together for promotion due to business considerations. We have to consider the stance and opinions of different manufacturers when associating various goods. This can be a limitation to the application of the data mined.Q:
Although 81's implementation has shown to be useful for the industry,
it seems that there have been quite a few di而 cultiesto apply the use of 81 extensively. Despite 50,
do you think it is still worth sparing resources for it in the future? A: We believe 81 in the retailing industry is still at the starting point,
but it is gaining companies' attention in the recent years. Global supermarkets and merchandisers such as Tesco and Carrefour have successfully integrated data mining into their businesses because they have recognized the business needs. The strong competition in e仟éct drives them to explore data mining. As the margin for this industry is thin,
retailers have to explore every method to maximize sales opportunities. 1 believe e-commerce marketing and supply chain can be further explored. Store design and planning can also benefit from 81 initiatives.Moreover, 1 am confident that 81 will have a bright future because the data involved in retailing industry is of such a staggering amount and so closely related to our daily lives. Customers can benefit more when they can buy suitable products with the right offers at store level. Even though we may encounter difficulties in implementing 81 in supermarkets' operations
,
the data collected will show its value to the business sooner or later.大型零售店數據分析師專訪
商業智能在超市業的應用
問:超市投入多少資源於分析部門? 答:雖然超市行業投放於分析部門的資源較銀行業少，但 隨著收集到的數據越來越多，過去幾年這個部門在超市業 內已得到認同，其重要性漸受肯定。 問:超市如何收集數據進行分析? 答:超市可以從會員暗的積分獎勵計劃、銷售點終端 (POS) 或於結帳時以及網上購物等多種渠道收集數 蟻。其中，透過網上購物這個渠道，超市可以追路客戶所 瀏覽的網站，並觀察他們的購買模式，例如對哪些產品感 興趣等。其他國家的超市甚至嘗試從閉路電視的片段分析 及研究客戶的行為，不過這種做法在香港尚未普及。 在銷 售點終端收集數擴是最常見的做法，因為在結照時收集到 的數據是現成的，而且能直接反映銷售的收入和利潤。從 會員暗的積分獎勵計劃取得的信息在將來使用時會非常有 用。從這些不同方式收集到的數據都可以幫助我們了解客 戶的購物模式。 問:超級市場會否將收集到的數據分類? 答:超級市場每天收集到大量不同類型的數撮，但並沒有 一套明確的分類方法。其中某些類型的數據，例如時間序 列，會用於編製專門報告。電腦系統會定期(每週、按 年/按月/連續 12個月)將收集到的數據製成報告，用於 評估公司的業務表現及運作是否健全。 間:超級市場一般較著重哪些指標? 答:超級市場的不同部門會參考不同的指標。譬如交易部 門會以各種貨品每天的銷售情況作為首要指標。銷售規劃 部門則會考慮每平古呎的銷售額數據，而負責管理供應鏈 的部門會留意現有的庫存資料。Interview
問:除了定期報告所載的特定數據外，超級市場 如何分析其他數據?有沒有相關的例子? 答:超市會不時指派團隊分析其他數據或將這項工作夕|、 判。這是因為我們現時既有的資料探勘技街仍不夠深 入，將數據外判給專家分析可以幫助我們發掘出更多更具 意義的結果。近數年，網上購物及電子商買的興起帶來 7 更多的數搶。而新的渠道如 QR code與傳統營銷手法互相 結合，也有助於深入瞭解客戶的需要。預計這種趨勢將會 延續下去。 問:超級市場將資料探勘應用於日常運作時，一 般會遇到甚麼困難? 答:由於我們有不少一直沿用至今的系統和軟件都是在較 早期設計的，所以要將資料探勘的新軟件有效地套用到現 時既有的系統上十分困難，因為新軟件與既有的運作方式 在設定上很容易出現不一致的情況。 之前提到，超級市場可以輕易通過日常銷售、長期顧客的 推廣計劃等多種途徑收集到大量原始數讀。由於早年的資 料庫在設計時只考慮到將這些資料應用到普通的推廣及宣 傳用途，或用作計算日常營運中的主要績效指標等簡單用 途，而商業智能是近年才開始逐漸受到重視，所以要將現 有的系統調較至適合進行資料探勘的程度，有一定的困 難。 若我們想在資料探勘方面再作進一步發展，現時數揚的儲 存方法如資料庫的設計、架構，以及一些正在使用的資料 庫管理系統，可能都需要一一修改設計或改良，甚至重新 建構。這不僅需要投入大量資源，還要承受在過渡期間因 日常運作受到影響而帶來的損失。 而且，進一步推進資料探勘的應用亦意昧著要更深入分析 顧客資料。不論這項工作是由我們的內部研究隊伍負責還 是外判給其他專家，在處理這些資料時，特別是依攝研究 結果作出商業決定的時候，我們都需要注意私障權的問 題，因為這很容易引起大眾關注。 除此之外，有時候即使我們成功透過資料探勘在大量數據 中找出陪藏其中的特殊關聯，但基於種種商業考慮，我們 未必能有效利用辛苦得來的成果。例如在外國有名的「啤 酒和尿布 J 的例子，即使找出 7 兩者的關聯，也不等於可 以將兩種貨品放在一起推廣，因為必需考慮不同品牌製造 商及批發商的意見。這無疑是資料探勘在商業應用上的一 大障礙。 問:雖然商業智能在零售業中有其價值，可是在 廣泛應用上似乎也有不少困難。你認為道在未來 仍是一個值得投資的項目嗎? 答:我相信，目前商業智能的應用仍只屬初步發展階 段，但已日漸受到重視。在國外，一些跨國大型超級市場 如樂購、家樂福等早已認清這方面的需要，並開始逐步將 資料探勘應用於營運之中，而且成效不俗。由於零售批發 業的利潤微薄，加上競爭激烈，因此商業智能應用的潛在 回報十分值得探索。這甚至乎可以在未來拓展至電子商務 的推廣，乃至整個供應鏈，當中蘊藏的商機亦是發展商業 智能應用的一大推動力。 再者，銷售業與我們日常生活息息相關，牽涉到的數據量 比任何一個行業都可能還要龐大，所以我相信在不久的將 來，這些數據都會在商業智能的應用下顯現價值，令各方 受惠。RM Digest
Algorithmic Trading
Technological advancement has brought about a shift to automated operations for many financial institutions. A good example is the rise of algorithmic trading
,
or algo trading in short.De自 nitionand development of algo trading
Algo trading is the kind of trading that makes use of computer to determine factors such as the asset to be traded
,
the trading amount and the trading method. The main difference between algo trading and traditional trading is that algo trading involves mainly computer operations instead of human decisions when making a deal. When developing an algo trading system,
at least one trading strategy will serve as a starting point. Rules will then be drawn up as guidelines to decide when to execute buy or sell orders as well as how to conduct the transaction,
such as by slicing a large order into smaller trades. According to a report from Aite Group in 2010,
there has been a surge in the use of algo trading across various asset markets,
especially in equities and futures.ln the U.s.,
the proportion of trades involving algo trading rose from 5% in 2004 to over 40% in 2010,
when more than half of the stock trading were done by computers automatically. There has also been a significant growth in the demand for algo trading in Hong Kong. Another research from Celent showed that the use of algo trading has increased to over 45% from about 25% in 2008.Advantages of algo trading
Nowadays
,
algo trading is widely used by hedge funds,
pension funds and other institutional investors. What triggered the shift in trading modes?The most obvious reason is the extremely short response time. When traders observe a trading opportunity
,
they need a few seconds to make the judgment. However,
in a fast-paced financial market,
such a delay may mean a great loss of profit. By using an algo trading system,
investment opportunities can be detected more easily and the deal will be completed in a few milliseconds. Algo trading can help achieve better investment performance and reduce the possibility of incurring a loss. Apart from the extremely short response time,
algo trading also enables financial institutions to identify the best trading strategy during the stage of system development. To determine the model and rules of an algo trading,
system developer may draw on traders' experience and apply technical analysis, index analysis or multi-factor analysis to real-time or histori臼 Idata. Since system developers do not have a well-defined percentage of distribution regarding the representation of the data,
software engineers will make use of historical market data as training data and input them into different systems to check for the system performance. This is why the quality of algo trading models is guaranteed.Since trading is done by machine
,
multiple and parallel transactions can benefit from algo trading. As such,
algo trading can be further applied to other trading techniques,
one of which is high frequency trading (HFT),
which has brought about a substantial change in market pattern. When placing an order thatinvolves a large amount
,
the price of the underlying asset will experience broad fluctuation and the trade will become less profitable. According to regulations of different exchanges,
transactions of this s日 le must be disc10sed to protect all investors, thereby exposing the trading strategies used by sizable institutions. HFT is a good solution to this problem. First of all,
an order will be divided into smaller parcels before they are sent to the exchange. Due to the low latency between the execution of each divided trade order,
all the divided orders will reach the server ofthe exchange at nearly the same time. The difference between the price of each of the deals is therefore minimal,
making it more difficult for other parties to trace the trading strategies and so putting institutional investors at a better position.R
M
D
i
gest
Besides
,
to hedge against market risk
,
traders would buy various financial instruments such as options and futures in addition
to trading in the spot marke
t.Making one transaction at each time will result in a higher cost and greater risks for financial
institutions. Algo trading enables them to control the cost of hedging more
e仟'ectivelyas trade orders are made nearly at the
sametime.
The risks of algo trading
As mentloned
,
algo tradlng systems are well
-
deslgned tools that help lower market risks
.
Nevertheless
,
there have been a lot
of concerns over algo trading. One argument Is that the algorithms are not stress-tested before they are put into use. For
ordinary programs
,
programmers try to come up with different testing environments to test the reliability of system. Testing
environments can also be set up for developing an algo trading system
,
but it is difficult to predict the factors contributing to
the extreme condition that leads to system failure. Therefore
,
it Is impossible to determine the tail risks of the system.
Moreover
,
compared with traditional trading methods
,
the speed of algo trading Is much faster.ln case the system encounters
errors
,
wrong tradlng orders would have been executed and enormous losses Incurred before the errors are belng discovered
and fixed.
A risk specific to the appllcation of algo trading Is the theft of trading model by developers or other staff members
,
slnce It
would be easier to extract trading strategies from the system than to elicit informatlon from traders. If someone steals the
model and sells it to other institutions
,
the system owner would
su仟erfrom the loss of revenue the system would be able to
generate
.
A more serious problem is that other Institutions may take action to Intercept thelr trades after obtaining their
trading strategies. On 3rd July 2009
,
a computer programmer of Goldman Sachs was arrested for belng suspected of stealing
trading programs.
Algo trading is most advantageous to institutional investors. But for other investors
,
It may be the cause of trouble if the
system is not well-designed. The flash crash In the U.S. market on 6th May 2010 was a concrete example of how poorly
deslgned system can cause huge fluctuation in the marke
t.The crash started in mldafternoon on that day
,
when the Dow
Jones Industrial Average Index plunged by about 600 points withln mlnutes and shedding as much as 998.50 points. It all
began with the sell-off of futures contracts by an automated trading system that triggered a large number of algo trading
software to execute stop-Ioss orders to close all related long positlons. Since the systems were not able to thlnk
Ii
ke a trader
,
they kept on selling even though the prlces of some stocks fell to virtually zero. Under specific conditions
,
algo trading can
have disastrous effec
t.The incident raised concerns from publlc and regulators regarding the reliability and potential market
risks arising from algo trading. The U
.s.
Securities and Exchange Commission has held a number of meetings discussing
possible regulatory measures on algo trading
,
but so far it has yet to reach a consensus.
Algo trading is becoming more and more popular and It Is expected that the trend will continue
.
ln order to protect investors
,
it Is high time for regulatory authorities to set up corresponding regulatlons and rlsk management measures to avold slmllar
incident from happenlng In the future.
算法交易
隨著科技進步，許多金國機構的營運操作漸趨自動化，算法交易 (algorithmic trading ，簡稱 algo trading) 的興起便是一個很好的
例子。 算法交易的定義和發展 算法交易指由電腦決定待交易資產、所涉金額及交易方式等因素的交易。算法交易 與傳統交易方式的主要分別在於算法交易執行時以電腦運算為主，以人為決定為 輔。在編寫算法交易系統時，首先會以最少一個交易策略作為起步點，接著定下規 則決定何時執行購買或出售指示，以及如何進行交易，例如把一個牽涉巨大金額的 交易分割成多個金額較小的交易。
RM Digest
根攝Aite Group 2010年的報告，算法交易在各種資產市場的運用激增，尤以股票和期貨為甚。在美國，涉及算法交易的交易宗數 比例由 2004年的 5%上升到 2010年超過40% 。於2010年，一半以上的股票交易都是通過電腦自動化進行。另外，算法交易在香港 的需求也有明顯的增長。 Celent的研究指出，在香港應用算法交易的水平在2008年時已達25%左右，現時更在45% 以上。 算法交易的優勢 算法交易現時獲機構投資者如對沖基金和退休基金等廣泛使用，是什麼原因觸發交易模式的轉變? 最明顯的原因是極短的反應時間。當交易員觀察到交易良機時，他們需要數秒作出決定。然而，在這個瞬息萬變的金融市場，即 使是些許的延遲也可能意昧著投資獲利滅少。使用算法交易系統可以更容易偵測到投資機會，並在數毫秒內完成交易。所以算法 交易不但有助提升投資表現，同時可降低投資損失的機會。 不僅如此，算法交易更可讓金融機構早於系統開發階段便能找出最佳的交易策略。系統開發者可以根撮交易員的經驗，以及對實 時或歷史數據進行技衛分析、指數分析和多因素分析，定出算法交易模型和規則。由於軟件工程師對資料的代表性並沒有一個明 確的分配比例，他們會使用過去的市場數攝作為訓練資料 (trainingdata) ，輸入到不同系統並測試其表現。故此算法交易模型的質 量能有所保證。 由於交易是透過機器進行，算法交易有利於重權交易和平行交易，因此可應用在更多交易技巧上，其中一項便是大大改變市場格局的高頻交易 (high frequency trading , HFT) 。如交易金額非常龐大，資產價格將出現大幅變化，交易所得的利潤亦會浦少。按
照不同交易所的規定，為了保障所有投資者，這類交易均要對外公佈，大型機構的交易策略因此便無所遁形。高頻交易是解決這 個問題的好方法。交易指令首先會被分成數個金額較小的交易，再把訊息發送到交易所。由於執行各個分割指令的延遲時間極 短，幾乎所有的分割指令都能同時到達交易所的系統，每筆交易的價格因此分別不大，其他人較難跟膝交易策略，為機構投資者 帶來優勢。 此外，為了對沖市場風險，交易員除了買進現貨資產外，也會購買不同的金融工具，如期權、期貨。每次只執行一個交易會令金 融機構成本及風險上升。但有了算法交易，不同市場的交易指令幾乎可以罔時發出，讓這些公司更有效地控制對沖成本。 算法交易的風險 如前所述，算法交易系統設計完善，有助浦低市場風險。然而，很多人對算法交易仍然有所顧慮。其中一項論點是算法交易未經 壓力測試便投入應用。對於一般的程式，程式員會嘗試構思不同的測試環境來檢測系統的可靠程度。算法交易亦有相應測試環 境，但無人能預知可能會引致系統故障的極端情況會包含甚麼環境因景，故此沒有人能判定該系統的尾部風險。此外，相對於傳 統的交易方式，算法交易的速度快得多。如果算法交易系統出現錯誤，不正確的交易指令可能在機構發現並修復問題前便已執 行，因而產生巨大損失。 交易模型被開發人員或其他員工盜取是運用算法交易的一個特殊風險。相比從交易員口中探取交易策略，從交易系統中套取交易 策略更為容易。如果有人得到模型並出售給其他機構，系統擁有者便會失去執行系統所獲得的收入。更嚴重的問題是其他機構在 獲得他們的交易策略後，可探取行動攔截交易。 2009年7 月 3 日，一名高盛程式員便因涉嫌竊取算法交易程式被捕。 算法交易為機構投資者帶來莫大幫助，但對其他投資者而言，如果系統設計不完善，可能會對金融市場造成大幅波動。 2010 年5 月 6 日美國股市閃電崩盤是一個好例子。當日下午交易市段中段，道瓊斯工業平均指數在數分鐘內下跌約 600點，最大跌幅 998.50 點。事件起源於某個自動交易系統拋售期貨合約。這觸發各個算法交易軟件敵動止蝕盤，將所有相關長倉平倉。由於系統本身不 能像交易員般進行思考，所以系統不斷拋售股票，其中有些股票的價格基本土跌至零。在特定情況下，算法交易可能會引致災難 性的影響。這次事件使公眾疑慮算法交易的可靠性和其引申的市場風險，亦引起監管機構的關注。美國誼券及交易委員會至今已 就算法交易的監管召開多次會議，不過仍然無法就此達成共識。 算法交易的使用日漸頻繁，預期這個發展趨勢將會持續。然而為了保障投資者，監管部門應盡快落實相應規範，並制定相關的風 險管理措施，以防止同類事件再次發生。
Business
Intelligence
Digest
Introduction to a common Bus
i
ness too
l (
PALO
)
What are 1181 tools"? How can they help Risk Management?
Generally speaking
,
Business intelligence (BI) tools are a type of application software designed to retrieve,
analyze and report/present data. The result is then used to make decision.To let you understand more about how BI tools work, here is an introduction to a common BI tool - PALO
,
written by one of our students:In the past three months
,
1 worked in an HSBC Insurancebranch as a summer intern.1 dealt with a lot of MS Excel 側的
everyday. In most business operations
,
we stored usefulinformation in numerous Excel spreadsheets
,
and usedExcel features such as pivot table to analyze data. However
,
there was too much business data
,
and they were usuallykept in separate Excel files. Do you think it difficult for the managers to make appropriate decisions from such Excel-chaos?
Palos
,
however,
has been popular these days. Palo is a commercial open source business intelligence solution for planning,
reporting and analysis. It has an Excel version,
which is called Palo for Excel. Although Excel already provides basic tools such as pivot table,
Palo helps fully utilize the potential of MS Excel and manage data in separate Excel 行 les at the same time,
and thus improvesproductivity and performance of a company.
Palo integrates data from each specific Excel file. There will be no more needs to have hundreds of sheets with often inconsistent data circulating throughout the company. The data is available to all users via the company's network
,
quickly and easily.Although Business Intelligence concept has become
popular among IT professionals in renowned companies
,
sometimes they Çlre discouraged to learn such business
solution tools. One of the reasons is that there are many vendors to choose: Oracle
,
SAp, SAS,
Microsoft,
IBM,
etc. Although big 5 one has powerful BI capability,
the cost is high and requires expert to deploy the tool. Some startupcompanies have no idea on what tools to use. Furthermore
,
companies may find it uneasy to find new blood who have experience on handling the BI tools. They need to input
resources and time in order to train their sta仟.
Palo for Excel
,
however,
is very convenient to use,
thus does not require too many resources for training. The only requirement for those inexperienced employees is knowledge of using Microsoft Office. Therefore,
Palo for Excel is easy to learn，的 ithas a familiar user interface. There are some other special features in Palo for Excel. The software does not save Excel tables but cell values only by using a comprehensive and indicative system of coordinates. This improves readability and f1exibility. Palo isalso web-enabled and not limited to the two
dimensions of Excel.
Palo also provides some formulae
,
which are also called PALO.DATA formulae; you can access cell values within the Palo database in an organized manner via intuitive object names. Data can not only be well displayed but alsocollected via Excel by overwriting the DATA formula with a
value. The formula will be preserved and will display the newvalue.
By combining the multi-dimensional and network-enabled
data power of Palo with the easy handling and graphic possibilities of Excel
,
we can now transform Excel into anunbeatable Business Intelligence tool for planning
,
analysis
,
reporting and dashboards.Palo's developer
,
Jedox AG has been increasing its licenseturnover
,
thanks to the success of Palo. The market share ofJedox AG should continue to grow.
Palo is an alternative if the company mainly focuses on
managing BI with Excel. Palo integrates the BI features into
MS office environment. It is especially suitable for SMEs as they may not have enough funding and human resources to buy and deploy BI tools and MS office is a common
application platform for every corporate.
