What is and how to use a data lake

Étiquettes

big data, computing, data lake, reactive programming

From storing huge information on economic parameters, market prices, customers behaviour, stress tests definitions and results, compliance legislations and rules, the financial sector and not only it tends to become a huge consumer of something that is called a data lake. Especially in a highly integrated risk management era and in a need for the fusion of multiple knowledge sources, when data pools are already there or available, data lakes are cost- and time-effective solutions coming in, both for data that you think you would like to analyse and for invading real time data.

Accordingly to its formal definition, a data lake is a storage repository that holds a vast amount of raw data in its native format. It means that data is not pre-categorized at the entry point and therefore, especially in online analytical processing, no optimal form is dictated by the fact that it has to support specific types of analysis. A data lake holds a vast amount of events.

The data lake solution provides a platform for a historical type of archive. It contains data from many different sources, with people in the organization being free to add or update data to the data lake. One launches Google type queries and then provides additional fields one may create and identify, searching interactively and expanding the description of the structure of the big data at the same time.

Its architecture seems to involve five components:

1. A double historical layer that gives information on all historical data. The batch layer has investigating services to search, locate, and access the historical data lake. The results are periodically re-computed and cached in a serving layer. One can use for example Hadoop to sift through the data and extract the chunks that answer the questions at hand, eventually replacing OLAP (online analytical processing).
2. A speed, on-the-fly layer that runs searching on updating, fresh data (say maximum one hour old), in real time, at low latency, eventually starting from a cached resulted field. It queues and streams the data, while updating the data lake, giving a view on the most recent data and favouring decision takings.
3. Lake services, which prepare, integrate and store the information from the data lake in loose pre-definable historical and on-the-fly double catalogues. With the searching results becoming available, new relationships are created between different sources of big data. Here data maybe safe and properly protected via tokenizing, encryption, key management and security audits.
4. A data reservoir, where you check the reliability and do the cleansing of the data, where people in the organization may access as necessary. There the data is prepared to answer specific questions. Through this type of container, big data actually becomes useful
5. A reactively managed engine to handle the real time constraints (an engine programmed to respond to the events, to scale to multiple cores and multiple server nodes, to be reselient to software, hardware and connection failures and to react in real time) that provides the libraries for the analyst experts to do their work.  With loosly coupled event handlers, as in reactive programming, the actual location of data looses importance from the tracing/functionality’s point of view and the data analysis gains in scalability and in event triggered response to the final user’s request.

The last layer actually represents the presently used software, with about three times more processing time being currently consumed by unautomated data sorting to make the data’s usage possible.

The data lake concept is about 5 years old, and there are not yet clearly differentiated vendors of complete data lake solutions.  Let them be  financial solutions or not Financial ones. Everybody is at the very beginning, but everybody is equally in the very need of competitive advantage. One visualises the financial data in the way that is understood and used at its best, as the V^3 approach (variety, velocity and vagueness of data) in operational risk.

Whatever the final goal is, clearly identifying what a data lake software concept might be about, in my opinion, a very good starting point.

Concrete elements of a Liquidity Contingency Plan

Étiquettes

example, liquidity contingency plan, liquidity risk

The liquidity management should include a liquidity contingency plan, scenario planning and testing of the plan. But little guidance can be found on what liquidity warning indicators should be monitored and what explicit strategic action should be tested. This is why I have summarised the points below.

Warning indicators

The purpose of early warning indicators is to alert management to the possibility of an impending liquidity crisis so that action can be taken quickly and early enough to avert it. Their monitoring may include:

• On the liability side:
  • unexpected and significant levels of withdrawals of retail deposits or non-renewal of wholesale funding facilities;
  • core retail deposit volumes falling below projected levels; and
  • a shortening of deposit maturities and a rise in requests to break fixed term deposits.
• On the asset side:
  • retail advances growing faster than projected;
  • a lengthening of loan maturities;
  • larger than expected drawdown of committed facilities;
  • a significant rise in undrawn committed facilities;
  • a rise in defaults and delinquencies; and
  • prepyments of loan facilities falling below historic behavioural norms.

Strategies to Test

The steps that can be taken to improve liquidity and on which one should have a sound idea on how the market reacts include:

• raising retail deposit interest rates;
• raising loan interest rates to discourage new borrowings and stabilise the balance sheet;
• usage of potential sources of funds;
• transferring liquidity to the affected group entity;
• capping balance sheet growth; increasing  on the offered products; and
• issuing of public statements (both locally and internationally) to deal with reputation risk.

The contingency funding plan should test:

• outright sales, or sales under repurchase agreements, of marketable assets;
• drawdown of committed facilities;
• funding from other group entities;
• contingent liquidity swaps.

Also, customers may withdraw fixed term or notice deposits under interest penalty and one should factor these balances in.

Are SEPA payments filling for the Euro-denominated LCR?

Étiquettes

EUR, government debt, LCR, optimize LCR, SEPA

It is not only that using payments-on-behalf-of (POBO) structures under the Single Euro Payments Area (SEPA), the companies can practically run their entire Eurozone operations through a single euro-denominated bank account. But SEPA has established a single clearing system for the 34 participating countries.

The Liquidity Coverage Ratio is defined as the ratio between available High Quality Liquid Assets and the net cash outflows. The denominator is the difference between expected cash outflows and expected cash inflows and has a minimum value of 25% of the total expected cash outflows.

Assume that, overall, the cash inflows are expected to be larger than the cash outflows for a viable financial institution and assume the institution will use a unique euro-denominated account to settle the SEPA transactions through a Pan-European Automated Clearing House as STEP2. At extreme, this means that the financial institution will have only euro-cash inflows from the clearing house (or at least close to zero euro-cash outflows), which will make the LCR mathematically ramp to very large values.

This might make the euro-denominated high quality liquid assets (usually government debt) useless from the LCR perspective.

Haircut risk to be born by the central counterparties clearing?

Étiquettes

credit risk, haircut, haircut risk, haircut risk definition, OTC, pricing, Variation Margin Gains Haircutting, VMGH

Basel strongly recommends Central Counterparties to be used, in order to manage the credit risk in Over The Counter contracts. According to the ISDA’s Variation Margin Gains Haircut (VMGH) mechanism (« CCP Loss Allocation at the End of the Waterfall »), it looks like this credit risk is going to be primarily covered by the Central Counterparty via  (haircuts in) the gains of the Clearing Members  which arise since the default is spotted. As the clearing is done via the Central Counterparty, there is always one of the two participants in the Over The Counter deal that is going to make a gain.

But this so-called Haircut (in the gains) Risk should show up somhow in the pricing of the OTC instrument, given that Over The Counter derivatives pricings are based on potential gains. Haircut risk looks like a type of credit risk, which appears conditional of the market (risk and gains), and might be triggered by and triggers itself liquidity risk, as the taken amount might be returned once the recovery done.  It is interesting how to price it, as it is a idiosyncratic risk (linked to the specific defaulting CM) which can be hedged, by making it market insenistive, with no future gains to show up.  But it turns out to be a hybrid with systemic risk, as it depends a lot on how many counterparties are defaulting, on average, within the Central Counterparty net.

Integrated risk modelling (credit, liquidity and market) becomes more and more necessary.

LCR impact on financial risk modelling

Étiquettes

credit risk, credit risk calculations, LCR, liquidity risk, market risk

“EBA FINAL draft Implementing Technical Standards on supervisory reporting under Regulation (EU) No 575/2013”, 26 July 2013, states that the Liquidity Coverage Ratio (LCR) has 31.03. 2014 as first reporting reference date, which is getting closer. But how should we get the reported values?

According to bcbs238 (“Basel III: The Liquidity Coverage Ratio and liquidity risk monitoring tools”, January 2013), one has to also look at the calculations of

• credit risk: in credit risk calculations the negative exposure (we own to the counterparty) is set to zero, as it does not have an impact if the counterparty defaults. However, via the LCR, this negative exposure has to be considered as net cash outflow:

120. Increased liquidity needs related to excess non-segregated collateral held by the bank that could contractually be called at any time by the counterparty: 100% of the non-segregated collateral that could contractually be recalled by the counterparty because the collateral is in excess of the counterparty’s current collateral requirements.

• market risk: if Option 1 is activated, we can borrow the missing amount of  High quality liquid assets (HQLA) at a fee, from a central bank. While the credit risk is technically not impacted by these new contracts, they will impact and generate market risk (interest rate and foreign exchange risks):

58. Option 1 – Contractual committed liquidity facilities from the relevant central bank, with a fee: For currencies that do not have sufficient HQLA, as determined by reference to the qualifying principles and criteria, Option 1 would allow banks to access contractual committed liquidity facilities provided by the relevant central bank (ie relevant given the currency in question) for a fee.

These two requirements imply an integrated credit, liquidity and market risk modelling in order to obtain consistent results.