最有看点的互联网金融门户

最有看点的互联网金融门户
传统金融的互联网化国际资讯

智能技术让流氓交易者无处可逃

传统金融的互联网化国际资讯

智能技术让流氓交易者无处可逃

2017年,伦敦同业拆借利率(LIBOR)和外汇交易丑闻等市场滥用事件推动全球贸易商监管支出攀升,总额高达约7.58亿美元。金融机构投入越来越多资源检测金融流氓行为,建设了包括监控交易、电子邮件和语音呼叫在内的很多系统,但由于系统彼此孤立,除非协同调查,否则无法查清交易者的完整行为。迫于预算压力,投资银行、资产管理公司和其他金融机构正在探索如何通过更有效的方式进行监督控制。

将自动化、机器学习和人工智能(AI)引入交易商监控技术范畴,或许可以为机构组织提供一个很好的创新机会。

有效监督是打击市场滥用的重要工具,但要完成有效监督难度很大,复杂性很高,正常行为和滥用行为之间的差别往往是很微妙的,合规性和前台团队需要筛除大量噪音。新技术可以通过识别模式和分析大量数据扰乱市场,金融机构面临巨大的麻烦,要彻底改变方式方法并制止资本市场的金融犯罪行之不易。

与监管同步

最近的Chartis和EY报告显示,超过70%的金融机构表示正在升级交易商监控系统,为最新的创新解决方案奠定了坚实的基础。监管机构对与交易相关的更详细上下文语境信息的要求越来越高,投入资金正是对这一要求的回应。同时,这也是一个机遇,遵守法规的目标仅仅是目标之一。交易活动的语境分析也正在成为最佳市场做法,金融机构因此可以更好地整合对传统电子通信和贸易数据,以及订单和电话语音数据等更广泛的数据来源的监测,提供真正全面的监测方法。

《欧盟金融工具市场法规(MiFID II)》的实施扩大了市场滥用监管的范围,许多金融机构在持续拓展其监管计划,以便于覆盖更多资产类别,如固定收益、商品期货和其他场外商品。然而,迄今为止,投资的趋势是使用传统技术跟上资产类别拓展和监管变化的步伐。在这些新的资产类别中,面对日益多样化的交易环境、特殊报告要求、惊人的数据属性(数量、种类和速度),管理股票和受监管的投资交易资产的系统正在苦苦挣扎。

在广阔的网络空间部署传统的基于规则的监控单体成本高而且局限性大,要对创新高效的技术建立信心是一个复杂的过程,需要大量时间并且取决于数据就绪状况。从基于规则的系统转变为基于模式的监控系统,金融机构必须权衡数据的复杂性和所有权,不同防线之间的治理难度以及与如何防御监管机构。

系统升级不会一蹴而就

鉴于上述情况的复杂性,规避风险的金融机构和监管机构不会选择立刻完成系统升级。逐步采取行动在效率和效率方面都可以做到收益可测,这在每个阶段都至关重要。如果金融机构能够制定从现有的、易于理解的逻辑系统到未来智能风险管理解决方案的过渡计划,就有机会向监管机构证明市场一体化在增强,并节省大量商业成本。关键在于向业务利益相关者和外部各方有力地评估和证明新技术在误报和漏报方面表现不断改善。

专注于内部系统和政策发展的公司应当把目光放得更加长远一些,合规要求不应当是唯一目标,在业务中应当创造更加有进取心、可持续的价值解决方案。以英国为例的基于原则的监管体系中,那些专注于满足最低合规要求的机构正在不断改造,以满足最新引导和预期。

创新机会

交易者监控系统目前每天发布成千上万则警报,导致监控团队的效率大大降低。实际上,金融机构增加合规团队人手以监控更多警报,成本也在上升。

然而,创新并不一定意味着从头开始。利用解析学和技术,可以使用新的数据源(如元数据、语音数据、社交网络数据、物理访问数据、财务数据、人力资源数据、生物识别数据和其他识别行为模式和重大风险的手段)整合和补充传统系统产生的孤立的交易数据和警报。

展望未来,关键在于应用智能技术提高效率,合规团队就可以专心完成深度数据驱动的调查,这样就可以保证监控资源降低成本的同时满足监管机构报告要求。金融机构现在认识到技术可以给业务带来的好处还不迟,使整个行业能够持续改进、可持续发展,既可以满足监管机构的要求又可以打击金融犯罪。不创新,则灭亡。

High-profile and costly market abuse incidents such as the LIBOR and foreign exchange (FX) trading scandals drove global expenditure on trader surveillance up to an estimated US$758 million in 2017, as financial institutions (FIs) continued to allocate increasing resources to detect and prevent rogue activity. As part of this response, FIs have deployed systems for monitoring trades, emails and voice calls, but they can often be siloed and may not tell the complete story of a trader's behaviour unless examined together. With budgets under pressure, investment banks, asset managers and other FIs are now asking whether the same, or better, surveillance controls are possible through more efficient means.

One such powerful opportunity for organizations to innovate is the incorporation of automation, machine learning and artificial intelligence (AI) into trader surveillance technologies.

Effective surveillance is a vital tool in the fight against market abuse, but it can be difficult to get right. It is a complex undertaking; the line between good behaviour and abuse is often nuanced and there's a great deal of noise for compliance and front-office teams to sift through. New technologies are disrupting the market with the ability to recognize patterns and analyze data from a growing number of sources. This is creating a perfect storm for FIs looking to overhaul their approach and disrupt financial crime in the capital markets.

Keeping pace with regulation

In a recent Chartis and EY report, more than 70% of FIs stated that they were in the process of upgrading their trader surveillance systems, laying strong foundations for the latest innovative solutions. The investment is primarily a response to regulators' demand for more detailed contextual information related to trades. However, this presents an opportunity beyond simply supporting compliance with regulations. Contextual analysis of trading activity is also emerging as good market practice. It can enable FIs to better integrate the monitoring of traditional electronic communications and trade data, as well as wider data sources such as orders and telephone voice data to provide a truly holistic approach to surveillance.

With the implementation of the Markets in Financial Instruments Directive (MiFID II) widening the reach of market abuse regulation, many FIs have been expanding their surveillance programs to cover a broader range of asset classes, such as fixed income, commodities and other over-the-counter products. However, to date, the tendency has been to invest in order to keep pace with asset class expansion and regulatory change using legacy technology. Across these new asset classes, systems originating from equities and regulated investment exchange assets are struggling under the weight of an increasingly diverse set of trading environments and idiosyncratic reporting requirements - as well as managing the sheer volume, variety and velocity of data involved.

Rolling out traditional rule-based surveillance siloes across this wider network is a high-cost endeavour with known limitations, and gaining confidence in new, more efficient techniques is complex, takes time and relies on data-readiness. When making the shift from a rules-based system to a pattern-based surveillance system, FIs must weigh-up data complexities and ownership, the difficulties of governance between different lines of defense, and defensibility with regulators.

Upgrading systems will not happen overnight

In light of these complexities, an overnight transition is unlikely to be an attractive option to risk-averse FIs and regulators. Making a gradual move that demonstrates measurable benefits in both effectiveness and efficiency is crucial at every stage. If FIs can build a transition plan from the current, well-understood logic systems of the past to intelligent risk management solutions of the future, they have an opportunity to demonstrate greater market integrity to the regulators, as well as making substantial cost reductions for the business. Key to this will be an ability to robustly measure and evidence performance improvements to business stakeholders and external parties across both false positives and false negatives of challenger techniques.

Firms focusing on the development of internal systems and policies should look beyond the compliance requirements and create more ambitious, sustainable solutions of value to their business. In a principle-based regulatory jurisdiction, such as the UK, those that remain focused on meeting the minimum compliance requirements constantly reinvent themselves to meet the latest guidance and expectations.

An opportunity to innovate

Trader surveillance systems currently generate thousands of alerts per day, drastically reducing the effectiveness of surveillance teams. Indeed, costs are rising as FIs expand their compliance teams to monitor the larger volume of alerts.

Innovation, however, does not have to mean starting from scratch. Through the use of analytics and technology, it is possible to consolidate and supplement disparate trading data and alerts from legacy systems with new sources of data - such as metadata, voice data, social networking data, physical access data, financial data, HR data, biometric data and other means of identifying behavioral patterns and material risks.

Looking to the future, the key will be to enhance efficiency by applying smart technologies so compliance teams can focus on deep data-driven investigations. This will ensure that surveillance resources reduce costs while maintaining confidence in meeting regulators' reporting requirements. Now is the time for FIs to recognize the benefits that technology can bring to their businesses, and to put the industry on an improved, sustainable footing that can both satisfy regulators and disrupt financial crime. The price of not innovating is simply too dear.

用微信扫描可以分享至好友和朋友圈

扫描二维码或搜索微信号“iweiyangx”
关注未央网官方微信公众号,获取互联网金融领域前沿资讯。

发表评论

发表评论

您的评论提交后会进行审核,审核通过的留言会展示在下方留言区域,请耐心等待。

评论

您的个人信息不会被公开,请放心填写! 标记为的是必填项

取消

栀航 | 未央团队未央编辑团队

66
总文章数

TA还没写个人介绍。。。

版权所有 © 清华大学五道口金融学院互联网金融实验室 | 京ICP备17044750号-1