Category Archives: Governance, Risk and Compliance
Earlier this week I met with security leaders at some of the largest organizations in the San Francisco Bay Area. They highlighted disturbing trends, in addition to the increased incidence of breaches they see increased:
- Numbers of customer who want to do security audits of their company
- Number of RFPs in which information is required about data security
- Litigation from data security breaches— and occurrences of class action lawsuits—as opposed to regulatory fines driving concerns
So much attention has been placed on defending the perimeter that many organizations feel they are in an arms race. Part of the problem is that it’s not clear how effective the firewalls are. While firewalls may be a part of the solution, organizations are increasingly looking at how to make their applications bulletproof and centralize controls. One of the high risk areas are systems where people have more access than they need to.
For example, many organizations have created copies of production environments for test, development and training purposes. As a result this data can be completely exposed and the confidential aspects are at risk of being leaked intentionally or unintentionally. I spoke to a customer a couple of weeks ago who had tried to change the email addresses in their test database. But they missed a few. As a result, during a test run, they sent their customers emails. Their customers called back and asked what was going on. That was when we started talking to them about a masking solution that would permanently mask the data in these environments. In this way they would have the best data to test with and all sensitive details obliterated.
Another high risk area is with certain users, for example cloud administrators, who have access to all data in the clear. As a result, the administrators have access to account numbers and social security numbers that they don’t need in order to do their jobs. Here, masking these values would enable them to still see the passwords they need to do their jobs. But it would prevent the breach of the other confidential data.
Going back to the concerns the security leaders had, how do you prove to your customers that you have data security? Especially, if it’s difficult to prove the effectiveness of a firewall? This is where reports on what data was masked and what it was masked to comes in. Yes, you can pay for cyberinsurance to cover your losses for when you have a breach. But wouldn’t it be better to prevent the breaches in the first place and showing how you’ve done it? Try looking at the problem from the inside—out.
While Dodd Frank received most of the media attention after the great financial crisis, during that period, the U.S. government signed into law the Foreign Account Tax Compliance Act (FATCA) back in March 2010 which will require Foreign Financial Institutions (FFIs) to report the names of U.S. persons and owners of companies who have bank accounts in foreign accounts for tax reporting and withholding purposes.
The law was set to go into effect on January 1, 2013 however on October 24, 2012, the U.S. Internal Revenue Service (IRS) announced a one year extension to January 1, 2014 to give FFIs more time implement procedures for meeting the FATCA reporting requirements. Banks who elect not to comply or fail to meet these deadlines will be tagged as a ‘non-participating FFI’ and subject to a 30% withholding tax on all U.S. sourced income paid to it by a U.S. financial institution. Ouch!!
The reasons for FATCA are fairly straight forward. The United States Internal Revenue Service (IRS) wants to collect its share of tax revenue from individuals who have financial accounts and assets in overseas banks. According to industry studies, it is estimated that of the seven million U.S. citizens and green card holders who live or work outside the U.S., less than seven percent file tax returns. Officially, the intention of FATCA is not to raise additional tax revenue but to trace its missing, non-compliant taxpayers and return them to the U.S. tax system. Once FATCA goes into effect, the IRS expects it will collect an additional $8.7 billion in tax revenue.
Satisfying FATCA reporting requirements will require banks to identify:
- Any customer who may have an existing U.S. tax status.
- Customers who hold a U.S. citizenship or green card.
- Country of birth and residency.
- U.S.-based addresses associated with accounts – incoming and outgoing payments.
- Customers who have re-occurring payments to the U.S. including electronic transfers and recipient banks located in the U.S.
- Customers who have payments coming from the U.S. to banks abroad.
- Customers with high balances across retail banking, wealth management, asset management, Investment and Commercial Banking business lines.
Although these requirements sound simple enough, there are many data challenges to overcome including:
- Access to account information from core banking systems, customer management and relationship systems, payment systems, databases and desktops across multiple lines of business which can range into the hundreds, if not thousands of individual data sources.
- Data varying in different formats and structures including unstructured documents such as scanned images, PDFs, etc.
- Data quality errors including:
- Incomplete records: Data that is missing or unusable from the source system or file yet required for FATCA identification.
- Non-conforming record types: Data that is available in a non-standard format that does not integrate with data from other systems.
- Inconsistent values: Data values that give conflicting information or have different definitions with similar values.
- Inaccuracy: Data that is incorrect or out of date.
- Duplicates: Data records or attributes are repeated.
- Lack of Integrity: Data that is missing or not referenced in any system.
Most modern core banking systems have built in data validation checks to ensure that the right values are entered. Unfortunately, many banks continue to operate 20-30 year-old systems, many of which were custom built and lack upstream validation capabilities. In many cases, these data errors arise when combining ‘like’ data and information from multiple systems. Given the number of data sources and the volume of data that banks deal with, it will be important for FFIs to have capable technology to expedite and accurately profile FATCA source data to identify errors at the source as well as errors that occur as data is being combined and transformed for reporting purposes.
Another data quality challenge facing FFI’s will be to identify unique account holders while dealing with the following data anomalies:
- Deciphering names across different language (山田太郎 vs. Taro Yamada)
- Use of Nicknames (e.g. John, Jonathan, Johnny)
- Concatenation (e.g. Mary Anne vs. Maryanne)
- Prefix / Suffix (e.g. MacDonald vs. McDonald)
- Spelling error (e.g. Potter vs. Porter)
- Typographical error (e.g. Beth vs. Beht)
- Transcription error (e.g. Hannah vs. Hamah)
- Localization (e.g. Stanislav Milosovich vs. Stan Milo)
- Phonetic variations (e.g. Edinburgh – Edinborough)
- Transliteration (e.g. Kang vs. Kwang)
Attempting to perform these intricate data validations and matching processes requires technology that is purposely built for this function. Specifically, identity matching and resolution technology that leverages proven probabilistic, deterministic and fuzzy matching algorithms against any data of any language, capable of processing large data sets in a timely manner and that is designed to be used by business analysts versus an IT developer. Most importantly, being able to deliver the end results into the bank’s FATCA reporting systems and applications where the business needs it most.
As I stated earlier, FATCA impacts both U.S. and non-U.S. banks and is as important for the U.S. tax collectors as well as to the health of the global financial and economic markets. Even with the extended deadlines, those who lack capable data quality management processes, policies, standards and enabling technologies to deal with these data quality issues must act now or face the penalties defined by Uncle Sam.
Data integrity is closely linked to the concept of trust which, in the world of human interactions, is based on a tight coupling between words and actions (do what you say and say what you do). In the IT world, this translates into first having a clear definition of data as well as how it is treated in the context of various business processes. If we have a clear definition of data, including policies such as access, privacy, change controls, etc. (the words), and if we have systems that consistently enforce the definition (the actions) then we have high trust and high data integrity. We know exactly what to expect, and the data always exactly matches our expectations. (more…)
Financial Stability Board Pushes Legal Entity Identifier to the G20– Vote Expected this Month – What’s Next?
Hot off the press! The Financial Stability Board (FSB) published today (June 8th, 2012) a report entitled “A Global Legal Entity Identifier for Financial Markets” for the G20 supervisors for consideration and response to the mandate issued by the G20 at the Cannes Summit for a final vote at the end of the month in Mexico. It sets out 35 recommendations for the development and implementation of the global LEI system. These recommendations are guided by a set of “High Level Principles” which outline the objectives that a global LEI system should meet.
The proposed global Legal Entity Identifier (LEI) is expected to help regulators identify unique counterparties across the financial system and monitor the impact of risky counterparties holding positions with the banks. Assuming LEI is approved by the G20 this month, it will be the first of these infrastructure standards to be implemented globally requiring firms to integrate, reconcile and cross-reference the new LEI with existing counterparty identifiers and information, as well as manage accurate and current legal hierarchies. (more…)
In a May 2012 survey by the Ponemon Institute, 66 percent said they are not confident their organization would be able to detect the loss or theft of sensitive personal information contained in systems operated by third parties, including cloud providers. In addition, the majority are not confident that their organization would be able detect the loss or theft of sensitive personal information in their company’s production environment.
Which aspect of data security for your cloud solution is most important?
1. Is it to protect the data in copies of production/cloud applications used for test or training purposes? For example, do you need to secure data in your Salesforce.com Sandbox?
2. Is it to protect the data so that a user will see data based on her/his role, privileges, location and data privacy rules?
3. Is it to protect the data before it gets to the cloud?
As compliance continues to drive people to action, compliance with contractual agreements, especially for the cloud infrastructure continues to drive investment. In addition, many organizations are supporting Salesforce.com as well as packaged solutions such as Oracle eBusiness, Peoplesoft, SAP, and Siebel.
Of the available data protection solutions, tokenization has been used and is well known for supporting PCI data and preserving the format and width of a table column. But because many tokenization solutions today require creating database views or changing application source code, it has been difficult for organizations to support packaged applications that don’t allow these changes. In addition, databases and applications take a measurable performance hit to process tokens.
What might work better is to dynamically tokenize data before it gets to the cloud. So there would be a transparent layer between the cloud and on-premise data integration that would replace the sensitive data with tokens. In this way, additional code to the application would not be required.
In the Ponemon survey, most said the best control is to dynamically mask sensitive information based on the user’s privilege level. After dynamically masking sensitive data, people said encrypting all sensitive information contained in the record is the best option.
The strange thing is that people recognize there is a problem but are not spending accordingly. In the same survey from Ponemon, 69% of organizations find it difficult to restrict user access to sensitive information in IT and business environments. However, only 33% say they have adequate budgets to invest in the necessary solutions to reduce the insider threat.
Is this an opportunity for you?
Hear Larry Ponemon discuss the survey results in more detail during a CSOonline.com/Computerworld webinar, Data Privacy Challenges and Solutions: Research Findings with Ponemon Institute, on Wednesday, June 13.
Today, agility and timely visibility are critical to the business. No wonder CIO.com, states that business intelligence (BI) will be the top technology priority for CIOs in 2012. However, is your data architecture agile enough to handle these exacting demands?
In his blog Top 10 Business Intelligence Predictions For 2012, Boris Evelson of Forrester Research, Inc., states that traditional BI approaches often fall short for the two following reasons (among many others):
- BI hasn’t fully empowered information workers, who still largely depend on IT
- BI platforms, tools and applications aren’t agile enough (more…)