K C Chakrabarty: Transforming credit information into action - issues and challenges
The views expressed in this speech are those of the speaker and not the view of the BIS.
Assistance provided by Shri Rajesh Jai Kanth and Ms. Jayanthi N. in preparation of this address is gratefully acknowledged.
1. Shri M.V.Nair, Chairman, CIBIL; Shri R.V.Verma, CMD, National Housing Bank; Shri M.V.Tanksale, Chief Executive, Indian Banks' Association; Shri Rajiv Sabharwal, Executive Director, ICICI Bank; Shri Arun Thukral, Managing Director, CIBIL; Shri Satish Pillai, Chief Operating Officer, CIBIL; delegates to the Conference; friends from the media fraternity; ladies and gentlemen! It is a pleasure for me to be present here this morning to deliver the keynote address at the Sixth Annual Credit Information Conference organized by CIBIL. It is heartening to note that CIBIL has completed 10 years of successful operations of its Consumer Bureau in India. While I congratulate the entire team of CIBIL on this occasion, I must also sound a word of caution amidst the euphoria that this occasion brings in. It is a time for CIBIL and indeed for all the credit information companies in the country to introspect and reflect on what the industry has achieved over the last decade and whither to from here? The spurt in the stressed assets in the banking system has put a spotlight on the effectiveness of the Credit Information Companies in the country. With huge increase in the data / information availability, could they have sounded a timely warning about the impending deterioration in the quality of assets in the banking system?
2. As the pioneer of the credit Information services in the country and as the most dominant player in the industry, CIBIL has a very critical role in how industry responds to the current challenges and adapts itself to remain relevant. In the financial services industry, the dominant player has to be much more resilient, much more responsive and much more restrained than others. These are the qualities you must develop. CIBIL has a tremendous responsibility, a responsibility to not only propagate the usage of credit information by the lenders but, at the same time, of ensuring that the consumers see merits in maintaining a clean credit history. I observe that the theme of the Conference "Transforming Credit Information into action: powering growth through data, analytics and decisioning" is very topical as it reflects a vision of taking the industry forward. In my address today, I would outline the essential pre-requisites of a good credit information system and how this information can be used for improving the efficiency of the credit delivery system in the country. I would also share my perspectives on some of the issues and challenges facing this sector and the regulatory / supervisory measures taken by RBI in this area. However, before I get into those issues, let me begin by giving a brief historical perspective of the functioning of Credit Information Companies in India.
Historical perspective
3. With a view to facilitating the distribution of credit to all sectors of the economy, RBI observed a need for establishing a Credit Information Bureau for the collection of credit information relating to borrowers fromlending institutions and for the provision of such information to the financial system. Accordingly, a Working Group (Chairman: N.H. Siddiqui) was set up in 1999 to explore the possibilities of setting up a Credit Information Bureau in India. The Group, in its report of November 1999, affirmed the urgent need for establishment of a credit bureau and accordingly, Credit Information Bureau (India) Ltd. (CIBIL) was incorporated in August 2000. CIBIL launched its credit bureau operations in April 2004 and its commercial bureau operations in May 2006. Following enactment of the Credit Information Companies (Regulation) Act (CICRA) in 2005, three other Credit Information Companies (CICs) have also been set up.
4. At the level of a regulator, credit registries can play a pivotal role in providing insights that can drive effective policy changes. They can also provide important inputs for the banking supervisors in monitoring systemic risks. A further use at a regulatory level may be to analyze appropriate capital and provisioning strategies for banks and, in particular, to assess whether current capital and provisioning regulations match up to actual risks. For the CICs/credit registries to be effective, the integrity and completeness of data is one of the essential pre-requisites. However, as I alluded to a little bit earlier, even after more than 10 years of their existence, the credit information companies/bureaus have not been able to penetrate the market in a meaningful way. There are a whole host of issues which have impeded the growth and acceptance of the CICs in India with quality of data submitted by the lenders, multiplicity of CICs, use of alternate data and consumer protection being the major ones. With a view to overcome these issues and challenges, RBI had set up a Committee (Chairman: Shri Aditya Puri, MD, HDFC Bank Limited) to examine the above issues and to also recommend a data format for furnishing of credit information to CICs by Credit Institutions. The Committee was also mandated to suggest best practices for the guidance of Credit Institutions in respect of usage of credit information as a part of their credit appraisal process. The final report of the Committee is awaited.
Information into action: the essential pre-requisites
5. Let me now turn to the theme of the Conference "Information into action." What do you need for putting 'information into action'? In the context of the Conference, this essentially means that available credit information has to be extensively used for drawing inferences for making lending decisions. Some of you might know that I have been a student of statistics; hence, I would use an analogy from that discipline to explain under what conditions can the information be used for decision making. In statistics, if any sample data is to be used to draw some inference about the population parameters, the estimates should satisfy four basic characteristics. These are "unbiasedness", "consistency", "efficiency" and "sufficiency". Similarly, if the credit information has to be used for facilitating effective lending decisions, the information should also satisfy the above four important characteristics. Let me explain them one by one.
(a) Unbiasedness: When I say that the information should be unbiased, it basically means that the information gathered should not be biased in favour or against a consumer, lender or a borrower. The CIC should not say that an inaccurate rating was arrived due to wrong data provided by a bank. The onus is on the CICs to ensure that the data that they gather is error-free and they are able to give unbiased information to their customers.
(b) Consistency: The credit worthiness of a borrower should be consistent and accurate. If one CIC rates a borrower as good and the other one rates him bad, this means that the credit information system is not consistent. Likewise, the rating of a borrower should not vary from one loan product to the other.
(c) Efficiency: An efficient information system should ensure that the CICs are able to provide reliable, accurate and unbiased information at the least possible cost.
(d) Sufficiency: The credit information system can be considered sufficient if it contains all required credit and payment related information about the borrower and no other CIC can give any additional material information on the credit history of that particular borrower.
Completeness of information
6. While I agree that at this stage we may not be close to meeting the four essential constituents of a robust information system, I would like to emphasize that unless we strive to meet them, we would not be able to transform information into action. A very important consideration for meaningful transformation of credit information into action is the completeness of data. Around 90% of our population has never borrowed from the formal financial system and hence does not have any credit history. Under the circumstances, they are quite likely to be denied credit by the lending institutions. How the completeness of data can be achieved in such cases? I will return to this aspect in a little while.
Quality of data
7. A central strand of completeness of information is the quality of data. The efficiency and effectiveness of a credit bureau is largely dependent on the quality of data collected in its database. RBI has issued guidelines for submission of timely, accurate and error free data to the bureau. However, our observation is that the data furnished by banks to CICs in many instances continue to be incomplete / inaccurate. As we all know, the quality of the product can only be as good or as bad as the inputs. The data containing junk values, incorrect information and incomplete identifier fields fed by the lenders to the CICs would only lead to half-baked outputs and eventually contribute to creating information asymmetries between borrowers and lenders. While, I understand that the quality of data held by microfinance institutions about their borrowers is comparatively better, the quality and consistency of data of the banking system is not very reliable and hence, poses significant challenges to the effectiveness of credit information system. The CICs in India are trying to address the issue by taking recourse to proven best practices and supporting technology to ensure that the credit information rendered by them are accurate, complete and updated. They have also started undertaking quality checks while receiving and uploading of members' data while returning the records rejected during the validation process back to the members along with reasons thereof, for correction and re-submission.
8. It is heartening to note that the results of efforts taken by CICs to clean the data have been positive as reflected in the improving 'hit rate' of the CICs. I understand that even in developed economies like the US, where credit bureaus have been in existence for over a century and where there are well evolved identification and verification systems, the bureau hit rate has been around 85%. So, while the hit rate of around 75% in the case of Indian CICs in the face of significant roadblocks like lack of unique identifiers and geographic and cultural diversities can be considered good, there definitely is scope for further improvement.
Alternate data sources
9. Let me now come to another strand of completeness. As I mentioned a little while back, a sizeable number of individuals/small business owners may not have a prior credit history. So, how the completeness of data can be ensured in such cases? Certain World Bank studies have indicated that inclusion of reporting of non-financial payment data (alternative data) proves extremely beneficial for making lending decisions, specifically for the segment not having access to credit. The financially excluded consumers who have a positive payment record in non-financial obligations, can have access to affordable credit. CICs have requested RBI to permit them to collaborate with sectors like Telecom, Utility and Trade credit for inclusion of their data in the credit bureau database. They have stated that inclusion of these data sources in the bureaus database will not only benefit financial sector but also be useful for the Telecom, Utility and Trade credit sector. The Committee on Comprehensive Financial Services for Small Business and Low Income Households (Dr. Nachiket Mor, Chairman, December 2013) has suggested inter alia, the need to develop a robust legal and regulatory framework around customer data generated in various transactions (credit and payments, digital and off-line), with the objective of customer ownership of their own transactions data and its use, among others, for signaling credit-worthiness. In this regard, the Committee has recommended that RBI should constitute a Working Group comprising of TRAI, CERC, and Credit Information Companies to develop a framework for sharing of data between telecom companies, electrical utilities, and credit bureaus. The suggestions/recommendations of the Committee are being presently examined by RBI.
Single v/s multiple CICs
10. There is a view that the existence of multiple credit bureaus in an economy ensures competition and creates a cost advantage for the users. However, where consumers as well as financial institutions do not have complete understanding of the benefits of bureaus, this could lead to disaster. Due to the functioning of multiple bureaus, only a fragmented view of the borrower can be obtained and a holistic financial picture may be hard to achieve. It has been suggested in several quarters that in an environment where such infrastructure does not exist at all, the initial impetus may have to come from the Government/regulator. In such a model, initially there could be a single CIC; and later on, a model with multiple players could evolve to avoid a monopolistic environment. In addition, besides setting up the bureaus, the government/regulator may also need to work towards education and awareness of both, the financial institutions as also the consumers.
11. In India, while setting up of CIBIL heralded the beginning of the credit information system about a decade ago, the Credit Information Companies (Regulation) Act, 2005 (CICRA) provided for an increase in the total number of CICs which may be granted the certificate of registration for carrying on such business. Thus, with the realisation that competition will bring better technology and efficiency in the system, three other companies were issued 'in-principle' approval to set up CICs in April 2009.
12. With at least 3 years of multi-CIC environment behind us, the problems of data fragmentation are coming to the fore. CICs collect credit information from banks and financial institutions, Non Banking Financial Companies, Housing Finance Companies, Micro Finance Institutions, credit card companies, etc. today. In terms of Section 15 of CICRA, every credit institution has to become a member of at least one CIC. Since four CICs are functional, there is the consequent issue of data existing in silos. As of now, when enquiry is made with one CIC about a certain borrower or prospective borrower, only the information that has been provided to the CIC by its own members would be available at that CIC. The information from the credit institutions from whom the borrower may have availed a credit facility but who are not members of that particular CIC, may not be included in the credit information report provided by it. This is another challenge with regard to the completeness of data.
Data on commercial accounts
13. The database in the CICs has been primarily used in the retail segment, especially while approving credit cards, personal loans, home loans and auto loans. Anectodotal evidence shows that a major chunk of this has been accounted for by the private banks and foreign banks with the public sector banks only recently starting to use them in a significant way. In the commercial segment, the use of the bureaus has not been to the same extent. SMEs are a big growth driver of any economy, accounting for a majority of jobs and exports. Inadequate credit flow has been one of the factors constraining the growth of the SME sector. One of the reasons for banks' reluctance to lend to this group is the unavailability of reliable information for banks to assess the creditworthiness of SME units. There is a significant role for the CICs in building an information base on SMEs and assisting banks in their credit decision making. This will facilitate the free flow of credit to SMEs and would provide an impetus to overall economic growth. The CICs would progressively need to enlarge their information database by bringing more sectors under their fold.
14. As I mentioned earlier, the Committee headed by Shri Aditya Puri is also looking into the issue of low usage of credit information by member institutions. We understand that some of the banks have complained about the high cost of obtaining a report as one of the considerations for low usage of credit information. On the contrary, the CICs are of the view that prices would come down with increased usage. In this context, the Committee has felt that workshops for creating awareness about Credit Information Reports (CIRs) and their use in credit appraisal by lenders would be more meaningful and these could be held regularly, in association with industry-level bodies like IBA or MFIN.
Unique identifiers
15. A study in August 2005 on "The Evolution of Credit bureaus in Asia-Pacific" had recognised that there are challenges relating to data quality and consumer identification issues that CICs in India (at that point only CIBIL was in existence) will need to overcome. Sadly, these challenges continue to plague the system. As banks in India have inadequate electronic data capture systems and inadequate networking facilities, problems encountered in data-collection may be magnified multi-fold. Notwithstanding the massive numbers of unique identities generated by the UIDAI, the country as a whole does not have a single official national identification registry for uniquely identifying an individual. On our part as banking supervisors, despite our continuous persuasion, the Indian banks have not been able to assign a unique identification number of their customers. CICs have tried to work around this problem by sometimes using the Ration Card number, Voters Id, PAN and Aadhaar Number or a combination of these. However, the results have much scope for improvement since mix-up of credit information results in customer grievance.
Customer grievance redressal mechanism
16. The Credit Reporting Knowledge guide recognises that "a challenge to data accuracy and validation is the lack of uniform identification schemes. Issuing national unique identity numbers is usually within the prerogative of the government. Adopting an identification system at the initial phases of establishing a credit reporting system would be ideal, but is not always realistic." In such a scenario, both Credit Institutions and CICs must have customer friendly policies in place. Credit institutions should ensure that the records submitted to CICs are updated regularly. Customer grievance redressal should be given top priority especially in respect of complaints relating to updation/alteration of credit information. CICs, at their end, should have a structured and systemic process for redressing grievances. They should abide by the period stipulated under the CICRA in respect of updation, alteration of credit information, resolution of disputes etc.
Summing up
17. I quite agree that there are significant hurdles to achieving completeness of information in terms of meeting the four fundamental criteria of unbiasedness, consistency, efficiency and sufficiency that I mentioned in the beginning. Some of the deficiencies of completeness of data/information can be overcome through use of better analytical tools by the CICs. In this regard, CIBIL could leverage upon its association with global leaders like TransUnion for providing much better solutions to their member institutions. A robust customer redressal system would also go a long way in improving the quality of information, by making the lenders aware about the importance of providing accurate data. The CICs need to be sensitive to the grievances of the customers and provide redressal in a timely manner. This would not only improve the quality of data but also lend credence to the utility of the credit information system in the country.
18. Notwithstanding the data quality and completeness issues, the lending institution should start using these information into their decision making process as unless the institutions use the data/information from the CICs in their decision making process and for pricing of their loan products, the completeness of information cannot be expected to improve.
Sharing of information and the individual's privacy rights / sharing of personal information
19. Having dwelt upon the essential pre-requisites for transformation of information into action let me highlight the issues of protecting the individuals' privacy rights insofar as CICs business is concerned. As per FSLRC report, personal information means any information that relates to a person or permits the inference of a person's identity. This includes demographic information such as the person's name and contact information; biometric information; and transactional information about holdings and dealings in financial products and services. The draft Code provides for the following protections for personal information:
- A prohibition on collection of personal information beyond what is required for providing the relevant financial service.
- A requirement to maintain confidentiality of personal information, unless the consumer has consented to the disclosure or it is otherwise permitted by law.
- An obligation to maintain accurate, up-to-date and complete personal information, and to allow consumers reasonable access to their personal information.
- Powers given to the regulator to specify additional requirements, exempt some financial service providers from application of this protection and establish mechanisms to ensure that consumers can access their personal information
20. The activities of credit bureaus are regulated almost everywhere so as to prevent violation of privacy and civil liberties. Privacy laws effect a wide range of issues that affect the individual, viz., limits on access to credit information by specified users (i.e. credit institutions and other entities allowed to use such information), setting time limits for how long such information can be maintained at the CICs, and the right to access, check and correct one's own credit information.
21. In India, the various provisions of CICRA cover the regulation of credit information business and also include responsibilities of the bureaus, rights and obligations of the credit institutions and safeguarding privacy rights of the individual. We can understand credit information about an individual at the CICs as a reflection of how he or she behaves with relation to his lender/s. If he/she is a good borrower, there is a record of such repayments; else a record of defaults is built up. This is akin to reputational collateral which is used by CICs to provide a picture of the individual to a prospective lender. In some cases, existing lenders themselves may like to review their portfolio to evaluate the functioning of such individual borrowers post disbursal of credits and also determine if the payment-performances have taken a turn for the worse. It may be emphasised here that the credit information report which is built up from such stored credit information does not by itself say whether the borrower or the prospective borrower is a good credit risk or bad credit risk. It only provides lenders with the data which allows them to take the decision themselves, depending on their risk-reward trade-off and their internal policies. This gives them the comfort of taking an informed decision based on the credit information available within the system.
22. It is not uncommon for questions such as "Why should my information be passed around in the system?" to be raised, even today in our context. The provisions of CICRA seek to strike a balance between the dual objective of putting in place an efficient system of sharing of credit information in the larger interest of financial system on one hand, and maintaining the privacy of such information, on the other. As part of their customer education and awareness efforts, the CICs should be able to effectively communicate the benefits of lower interest rates that the consumers may derive if they are perceived as 'good borrowers'.
Recent regulatory / supervisory measures
Let me now turn my attention to few of the recent regulatory/supervisory measures we have initiated in this area.
(i) Central Repository of Information on Large Credits(CRILC)
23. As you all know, Dr. Raghuram Rajan on assuming charge as Governor, RBI announced that RBI proposed to collect credit data and examine large common exposures across banks which will enable the creation of a central repository on large credits, which may be shared with the banks. This is expected to enable banks themselves to be aware of building leverage and common exposures. The CRILC is expected to act as a repository of large credits containing system-wide exposure of individuals and entities having exposure (both fund and non-fund based) of more than Rs. 50 million. Thus, the basic purpose of CRILC is to collect, store and disseminate large credit data to banks. The underlying objective of the creation of the CRILC is to tackle the menace of rising NPAs in the Indian banking system. I would like to urge the banks and the CICs to also put in place a system to collect credit data of exposures which are even below Rs.50 million, so that they could have reliable information about all the borrowers in the financial system.
(ii) Investment in credit information companies
24. RBI has issued a revised directive on November 29, 2013, allowing higher FDI limits to entities which have an established track record of running a Credit Information Bureau in a well regulated environment. This regulatory measure is aimed not only at bringing in the much needed foreign equity support but also to keep intact competition in this sector. As a result of this measure, the foreign partners of the CICs are taking steps to increase their equity stakes in the CICs and thereby continue to bring in foreign technology support to this long gestation capital intensive industry.
Conclusion
25. Before I conclude, let me revisit the reasons for why the credit information bureaus including CIBIL, have not received the level of success they would have expected, even after a decade of their existence. In essence, the two users of the credit information, the lending institutions as well the consumers are still wary of the outputs generated by the CICs. While on the one hand, the lenders do not see compelling reasons to subscribe to the CIC reports for want of appropriate information or analytics, the consumers on the other hand are still sceptical about the authenticity of the CIRs. There have been numerous complaints about the CICs handing out incorrect reports to the consumers who have to then run from pillar to post to get these corrected. The grievance redressal system in this area, in fact, needs a thorough overhaul. Thus, going forward, the challenge for the CICs and indeed, for CIBIL, which has a virtual monopoly over this business is to address the twin concerns that I have mentioned. The purpose of having CICs in the system would be best served when the credit needs of deserving people are met in a timely manner and at a reasonable cost.
26. I am sure that the Conference would throw up some 'actionable' ideas which would go a long way in not only spreading awareness about the credit information system in the country but also improving it. I once again congratulate CIBIL for reaching a significant milestone in their journey and thank them for inviting me to be a part of this Conference. I wish you all a very fruitful deliberation ahead.
Thank you!
References:
The Evolution of Credit Bureaus in Asia-Pacific- Dr. Rita Chakravarti and Beng-Hai Chea (August 2005)
Role and Effects of Credit Information Sharing- - Tulli Jappelli and Marco Pagano (April 2005)
Credit Reporting Knowledge Guide, International Finance Corporation (2012)