Main Content

How Smart Resolution Planning Can Help Banks Improve Transparency, Increase Profitability and Reduce Risk

Resolution and Recovery Planning (RRP or “living wills”) is a new regulatory model, requiring new data, new processes, new tools, and new thinking. 

By Paul Lippe

Resolution and Recovery Planning (RRP or “living wills”) is a new regulatory model, requiring new data, new processes, new tools, and new thinking, just as the Securities Exchange Act back in 1934 helped create modern accounting standards and reporting models. When undertaken effectively, RRP can help banks manage transparency regardless of scale using a data-centric approach, modernizing the legal function, and giving banks the benefits of superior analytics.

The disciplines needed for resolution planning are consistent with pricing risk and disposing of assets. Lacking such granular understanding, banks will struggle to adhere to their risk strategy and appetites, which is not acceptable in a world of tighter margins and new competitors.

Given their novelty, it is not surprising that the first wave of living wills have fallen short of regulators’ expectations. According to Thomas Hoenig, the vice chairman of the FDIC, earlier-filed plans “provide no credible or clear path through bankruptcy...”

But at the same time, Hoenig, reflecting on the FDIC’s experience with speedy resolution for smaller institutions over a “resolution weekend,” pointed a way forward, saying banks that “prepare acquisition and restructuring plans for clients, corporations, and financial companies across the globe” should be able to devise plans to meet RRP requirements

As Hoenig suggests, by applying mergers & acquisitions (M&A) best practices, banks can use RRP to improve operational efficiency, risk management, and profitability. Since August 2014 there have been several developments to help break the RRP logjam:

  • In April, the Treasury Department’s Office of Financial Research (OFR) published “Contract as Automation: The Computational Representation of Financial Agreements” describing a model for creating easier to understand contracts;
  • Banks have increased their focus on operational excellence and analytics to respond to new market forces; 
  • Several of the authors (and other innovative companies) have been collaborating to bring new tools and methods to bear on the complexity challenges of RRP, integrating the latest in “design thinking” from Stanford ( and elsewhere. 

The next wave of living wills presents an opportunity to establish superior practices to move regulators and systemically important financial institutions (SIFIs) forward in the implementation of RRP.

Absent success in resolution planning, politicians and regulators will continue to advocate more stringent size and leverage limits, and will continue to push hard on banks and enact penalties for non-compliance. Smart resolution planning is far and away one of the most cost-effective compliance investments for banks.

RRP Can Have a Positive Impact on ROE

There are five areas where SIFIs can improve their financial performance with the improved transparency that will come out of RRPs:

Risk assessment and pricing of assets, especially derivatives – Financial assets are defined by complex agreements, which typically contain unstructured data that is very difficult to evaluate. Incomplete understanding of complex derivatives remains a significant threat to banks’ ability to restructure with confidence. While a bank may feel confident in its risk management of a derivatives book on a “business as usual” basis, it may not feel so confident in its management of that book against the backdrop of significant asset disposals and other restructuring transactions. The development of better information systems and the corresponding integration of previously disparate legal information will help banks better manage, price, and hedge risks.

Capital levels – National capital levels are set in part based on assets and in part on inter-company obligations. Better inter-company agreements and a much more precise understanding of assets will help prevent tying up more capital than is necessary, providing banks with a better understanding of their most capital-hungry assets.

Asset dispositions – As we saw in April when one of the world’s largest manufacturers announced its disposition of $160 billion of assets, global banks will regularly sell and buy assets in response to pressures from both regulators and investors. Because buyers have similar information and risk aversion problems as sellers, having well-structured information that buyers can quickly analyze and get through their due diligence process should increase the chance of multiple bidders emerging and improve the quality of bids, boosting sale prices.

Liquidity levels – New bank liquidity requirements depend primarily on a precise understanding of a bank’s contractual rights and obligations. A more exact, real-time understanding of these requirements will help a bank better manage its costly liquid asset maintenance requirements. It may also assist banks to model their liquidity requirements more effectively through a better understanding of the circumstances in which liquidity outflows may, or may not, be expected to take place. 

Procurement – In order to operate, banks buy an enormous amount of non-financial goods and services that are described in complex agreements and practices, which must be better understood for continuity of services. Both buyers (banks) and sellers can gain a better understanding of these agreements and practices through the information that is collected as part of the RRP process. Banks can develop better information for compliance reasons. Buyers and sellers can understand and modify contracts to adapt to banks’ regulatory requirements (including provisions in contracts addressing expressly what happens on resolution) and dealing with other important requirements. 

What’s New (and What’s Not)
The Transparency Gap
The single biggest factor that has changed in banking and financial markets (and the world at large) is transparency. Much more information is much more widely available. Many of the recent enforcement actions (e.g. Libor, London Whale, forex manipulation) have come about because information systems, including email and chatrooms, capture comments that would probably have faded into the ether in the pre-digital world. Similarly, trading and lending are transforming as more players have more access to more information.

But contra the broad trend toward transparency, the financial crisis and the RRP process revealed that SIFIs do not particularly have a mastery of their data, management systems, nor the contracts that define their assets and liabilities to facilitate compliance with RRP. This reflects choices, habits, and interests that have developed over time: 

  • Most SIFIs are an amalgam of predecessor entities with discrete systems and operating cultures that have never been fully integrated.
  • Banks are run as a group of P&Ls that cut across entities, structures, and contracts, so the primary analytic and reporting focus is on the P&L that drives compensation metrics. A particular P&L may have little to do with legal entity structure and will be the responsibility of divisional management that may, again, have little correspondence with legal entity management structures such as boards of directors.
  • Legal work product has traditionally been created on a one-off basis that was not designed for reuse; on paper, in Microsoft Word, and occasionally in Excel, but almost never as structured data. Today’s reality is quite a long way from the future world of “computational contracts.” This lack of structure makes it notably harder to apply technology solutions to legal work product than to other types of information.
  • Even when databases are used to store legal work product, it has been historically difficult or impossible to query that database for substance. Searching legal documents according to the criteria that matter for RRP (e.g. events of default, assignability) typically requires attorneys undertaking a labor-intensive due diligence exercise. 

Similarly, it is clear that under the Bank Recovery and Resolution Directive, which sets out the E.U.’s version of RRP, bank prudential regulators are looking for bank entities to demonstrate a deeper understanding of their critical functions.


To be a global, multi-business line bank is to be complex. Unlike most international businesses, which are highly integrated, even across geographies, global banks are highly federated, with hundreds of distributed P&Ls that cut across regulated geographies and legal entities. Correspondent banking practices create complex cross-border booking models that only add to the challenge of understanding how risks are originated and where they reside.

The global bank structure, combined with increasing national regulation, inevitably gives rise to tensions that show the need for a “new normal” legal operating model:

  • Localized P&Ls take risks, some of which turn out badly in unpredictable ways, leading to both unpredicted losses and unstructured “risk aversion.”
  • National regulation means higher local capital requirements and varied operating rules, including formal or informal ring-fencing and mandated independence of local management, making it difficult to achieve economies of scale, and to impose global operating policies across a banking group.
  • An Anglo-American legal and regulatory culture often conflicts with the acceptable norms in parts of Continental Europe, Asia, and Latin America.
  • “Old normal” legal functions are reasoning-based and subjective, not information-based and objective, so they tend not to contribute to improving systemic insights or quantifying risk. This can result in under- or over-pricing of risk, because the local P&Ls tend to under-price risk and the central functions tend to over-price risk.
The combination of complexity and higher capital levels can be a significant headwind to increased profitability, with return on average equity in the U.S. hovering at only around 10 percent, compared to 15 percent from the 10 years ending in 2004. Even without the regulatory pressure to understand and simplify their structures, banks face investor pressure to reduce complexity.

Disruptive Competition 
Long before Harvard Business School Professor Clayton Christensen coined the terms “Innovator’s Dilemma” and “Disruptive Innovation,” savvy managers were afraid of new, lower cost, more technology-enabled competitors. Global banks now face two classes of disruptive competitors: greenfield, lower cost start-ups like Lending Club, OnDeck, and Virgin Money, and established customer-connected companies such as Google and Apple.

To simultaneously confront an intensive, costly, and friction-creating regulatory regime and new disruptive competitors is not a comfortable place. Many companies have lost out on market opportunities when they’ve been confronted with the twin challenges of increased regulation and new, lower-cost competitors. Simply pushing the anti-regulatory rock up the hill is likely to be a Sisyphean undertaking.

Role of Legal Work

Paper and ink was certainly the pre-eminent technology for two millennia. Today digitally stored and managed information is much easier to access, understand, and manipulate. One lesson of the financial crisis was that many risks went undetected for lack of systematized management of key legal information, as sorting through problems was quite cumbersome. 

To meet the challenge of RRP, lawyers will need to change how they deliver legal work so that it becomes more systematic and standards based. 

Professor Larry Lessig of Harvard Law School has famously written that “code is law.” In the same way, we can say that “contracts and other legal work product want [indeed need] to be code” in order to be readily translated into business values like risk, asset value, optionality, and cash flows. Recognizing the slowness of lawyers in adopting new work styles, and the vast number of contracts that already exist, it will be quite some time before we could imagine most contracts having been written in a computational style. But taking advantage of new technologies we can restate existing contracts and other unstructured legal product to achieve most of the potential future benefits of computational contracts in response to the RRP challenge.

RRP – The Path Forward

A living will is effectively a roadmap and simulation of the largest possible series of transactions in a bank’s lifetime. This type of analytical exercise is common in electronic systems design or software testing, but unprecedented in law and finance. It is axiomatic across the technology industry that to manage complexity requires standardization, modularization, and simplification – not exactly the attributes one could use to describe the design approach used by the legal department at many firms.

Even if we can master the techniques of managing complexity, we should acknowledge that perfect transparency is often uncomfortable. In any part of life, there is a certain cultural and emotional dissonance in being quite so transparent about what one is likely to do in changing scenarios. Nonetheless, just as a will is prudent practice for wealth management, even though death and disposition of assets to one’s heirs is hard to talk about, so the living will is a prudent step. 

It takes no great insight to recognize that given the political climate, governments will scrutinize most banks’ plans until superior practices are established. Ultimately, regulators will expect banks to use tools and methods that are a match for the complexity of the problem. 

Here’s an overview of the flow for best-in-class RRP:

  1. Capture or “grab” existing contracts. The source documents are typically electronic image files (e.g., PDF, other format) and may not be in a searchable format. Older documents may be poor-quality image files, but must still be reviewed.
  2. Contracts are initially checked for completeness, a folder structure for the contract family is created, and documents are uploaded into a secure online collaboration system. Data on the number of contracts, the number of contract families, escalations, and similar metrics are captured. This process can be automated with machine learning.
  3. Contracts are converted into searchable HTML and further checked for missing pages, documents, or other errors. Common or hierarchical terms can be passed from “parent-child” or to “siblings.”
  4. Contracts are analyzed according to pre-defined data points, precise questions that yield yes or no answers. Questions and problems are resolved through a flag and escalation process. 
  5. Multiple types of reports are generated. Management information reports identify contracts and responses according to pre-defined criteria. Online reports support drilling-down to the specific page of a contract. The report generator allows very specific criteria, e.g. date range, to be specified for report output.
  6. Contracts are grouped by family, so, for example, multiple contracts with the same counter-party are linked and consolidated (or not) as appropriate. Because the information is in a database, anybody seeking information can run a report to see the relationship between different entities and agreements. A large SIFI might have 600 different agreements with a major supplier or counter-party. 

It takes no great insight to recognize that given the political climate, governments will scrutinize most banks’ plans until superior practices are established. Ultimately, regulators will expect banks to use tools and methods that are a match for the complexity of the problem. 

Now Enter IBM Watson 

Many readers will be familiar with IBM’s Watson, the first commercially available cognitive computing capability, ushering in what IBM CEO Ginni Rometty has called “a new era of computing.” Having tackled some of the largest contracts analytics problems, OnRamp joined the IBM Watson Ecosystem and is applying Watson as well as a range of other related machine learning and natural language processing technologies to:

  • Ingest and extract meaning from large volumes of text; 
  • Provide baseline meaning as needed for RRP, and ultimately connect to other systems for more sophisticated analytics; and
  • Allow professionals with appropriate experience to execute work with increased scale, speed, consistency, and quality.

Though most legal work product has some sort of structure, it is not particularly useful unless read by an expert lawyer, limiting a SIFI’s ability to understand its tens of thousands of agreements. Historically, latency – or what might be called the “open syntax” of the law – has frustrated attempts to automate legal tasks. 

Using these new tools and methods, we can now map thousands of documents quickly, with high quality and reasonable costs. By integrating people and technology, we can “sample and simulate” perhaps as few as 5,000 of a bank’s 100,000 documents, training a system on that subset so that it can, with appropriate human oversight, rapidly analyze, learn and discover new patterns and insights from the remaining documents. 

If we transform contracts to make them look and act more like computational objects in integrated information systems, legal information systems could “talk to” accounting systems, compliance systems, and other relevant corporate information systems. To enable this, legal work product must be converted from existing flat storage methods – such as image files, PDF, or Microsoft Word – to  databases, meta-information, and ultimately code. Like any complex change, this large-scale transformation wouldn’t happen in the ordinary course; without it, RRP is unlikely to deliver the compliance, de-risking, or return-on-equity benefits that it promises.

Banks can leverage the work required for resolution planning to provide insight and efficiency across many business processes, using RRP as the legislative authors intended to reduce risk and improve operational excellence. Longer term, contracts and other forms of legal work product will become more accessible and easy to manage. One can imagine that in future banks will find ways to form contracts more quickly, understand those that already exist, and share information and expertise across disciplines. This type of “cognitive backbone” could then be used to integrate forms of structured and unstructured data, simplify business processes, reduce risk, generate insights and empower employees. As organizations unlock the unstructured data in documents and integrate it with people and structured data, the opportunity goes well beyond cost savings: after all, expertise is the most important asset of an organization, yet the differentiation comes from how organizations empower their people.


Resolution Planning is a fundamental element of the regulatory framework developed to manage systemic risk. Fortunately, we have new tools to manage complexity, reduce risk and improve profitability. 

RRP is not just about compliance, but operational excellence, and the banks that succeed at RRP will face both fewer regulatory headwinds and lower operational risk and cost of capital. Although there will be many challenges, a thoughtful approach to living wills can provide far greater benefit than compliance with the mandates alone – and become a realized marketplace advantage.

Sidebar: The Cisco Legal Model

For more than a decade, the Cisco legal department has been known as a leader in bringing innovative approaches involving People, Process & Technology to law. 1

Cisco Senior Vice President and General Counsel Mark Chandler follows two management themes that are very different from conventional practices: “running legal like part of the business,” and being a “gateway, not gatekeeper.”

One of the areas of focus for Cisco is efficient M&A processes, as well as scaling and streamlining services delivery. As lawyers have moved from Cisco to SIFI banks, they have begun applying Cisco-based approaches to manage the large-scale legal complexity challenges of banks.

Sidebar: What is Cognitive Computing?

As we ask computers to understand and reason more like we do, we must teach them not just the structure and vocabulary of spoken and written languages but the words and concepts that are particular to professional and business domains. When they have that kind of capability, they can begin to offer us not just answers to questions but fresh insights that might not have occurred to us on our own.

Cognitive computing aims to understand and answer questions in the same way as humans communicate – often referred to as Natural Language Processing or Deep QA.

The capability to process large amounts of data from all kinds of internal and external sources, not bound by volume or memory – including structured data, unstructured content and even images – in order to interpret data to expose patterns and connections through both paraphrased and inferred information is a characteristic of cognitive computing. 

Cognitive technology, such as IBM Watson, is built to mirror the learning process that we have – through the power of cognition. What drives this process is a common cognitive framework that humans use to inform their decisions; Observe, Interpret, Evaluate and Decide. Yet instead of the classic human master-apprentice, Watson ingests the corpus of data, or collection of information, and is then trained by human experts to learn how to interpret the information.

Cognitive technology is revolutionizing the way we make decisions, become experts, and share expertise in different industries, and it is discovering and offering answers and patterns we hadn’t known existed, faster than any person or group of people ever could.

Source: Smart Machines: IBM’s Watson and the Era of Cognitive Computing (Columbia Business School Publishing) – October 2013