Workshop on Firms, Industry and SME Data in India Venue: ISI Delhi, New Delhi, India Date: March 27th, 2015, Time: 10:00am to 4:00pm Knowledge partners Sampling & Official Statistics Unit (SOSU) Indian Statistical Institute, Kolkata PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 An international workshop on firms, industry and SME data was organised jointly by the IGC (International Growth Center),IFMR Lead and Sampling & Official Statistics Unit (SOSU) at ISI Delhi on 27 March 2015. Data on Industrialization, Firms, Employment and MSME sectors are key sources to study the characteristics of a national economy, and more specifically - the Indian economy. However, the researchers using official statistics on Firms, Industry and MSME often face problems in conducting their studies due to the nature and unavailability of appropriate data in these sectors. Those issues remain mostly unreported to the official statisticians who are in charge of collecting the national datasets which are never assessed and resolved. This workshop was conducted to pursue the dialogue started on these issues but focused more on the type of economic data dedicated to Firms, Industry and SMEs in India. The central themes of the workshop are National Firms, Industry and SME data in India, comprising the Economic Census Survey, the Annual Survey of Industries and the NSSO Employment and Unemployment Survey. It will allow researchers from different horizons: academics, policy makers (UNIDO, etc.), Government agencies, etc., to exchange on the difficulties they face in using or collecting these datasets and define the steps of a way forward to improve the quality of the data available in these areas. Prior to the workshop a study will be conducted by a team of academics towards the researchers based in India and worldwide working on these sectors of the Indian economy. The objectives of this workshop are: 1) Understanding of the industry, firm and SME data environment; 2) Identify researchers needs and research topics yet to explore in these areas and the data issues related to them; 3) Initiate and carry on with the dialogue between researchers and official statisticians. PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 Participants: - DrPronabSen, IGC India Central; National Statistical Commission - Dr.Samuel Asher, Oxford University - Prof.. Marc-Andreas Muendler, University of California San Diego - Prof. Kala Krishna, Pennsylvania State University - Prof. JagadeepSivadasan, Stephen M. Ross School of Business, University of Michigan - Prof. B.N. Goldar, Institute of Economic Growth, University of Delhi - Shri G.C. Manna ,Additional Director General, Economic Statistics Division, MOSPI - ShriSanjibBasu, Director, Directorate General of Commercial Intelligence and Statistics - Dr. TCA Anant, Chief Statistician of India and Secretary MOSPI - Mrs Sharon Buteau, IFMR Executive Director - Dr. JyotiMukhopadhyay,Senior Faculty Associate (Economics Area), IFMR - Dr. AmeetMorjaria, Harvard University - Mr Stephen O’Connell, City University of New York - Dr.SandipMitra, Assistant professor, Sampling and Official Statistics Unit (SOSU), ISI Kolkata - Dr. AsitChakraborty, Visiting Scientist, Sampling and Official Statistics Unit (SOSU), ISI Kolkata, formerly with RBI - Dr. PrabirChaudhuri,Sampling and Official Statistics Unit (SOSU), ISI Kolkata, formerly with NSSO - MrinalBhowmik, Sampling and Official Statistics Unit (SOSU), ISI Kolkata, formerly with CSO - Dr. Urmila Chatterjee , World Bank Delhi - Dr. Dennis Rajakumar,EPW Director - Dr.Joseph Mariasingham, Statistician, Development Economics and Indicators Division, ADB Manilla - Mrs Johanna Boestel, ADB India - Mrs Kanupriya Gupta, ADB India - Mrs Nidhi Kapoor, ADB India - Mrs Radhicka Kapoor, ICRIER - Dr.Monika Goel - Mr Vikas Dimble, IGC India Central -Mrs Nalini Gulati, IGC India Central - Mrs Freeda D’Souza, IGC India Central - Mrs Pratibha Joshi, Associate Director, Small Enterprise Finance Centre - Mrs Aurelie Larquemin ,IFMR Policy and Development Fellow - Mr Tanay Balantrapu, IFMR Research Manager Elements shared with the audience beforehand: - The Use of Firm Data for Development Research Around the World, Implications for India, by Sam Asher and Paul Novosad, March 2015. PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 SESSION I – Sam Asher and Paul Novosad’s paper presentation and discussions Sam Asher presented thepaper “the use of firm data development research around the world” written with his co-author Dr Paul Novosad. He mentioned that there has been lately a boom in firm research due to a boom in the availability of firm data and the development of computerization. India would be the main beneficiary and has surely a lot to gain from people accessing its data. He spoke about the different users of data and different types of datasets:Census – exhaustive information (Government); Sample– e.g. NSS (Government/Private); Transaction – information on firm’s actions (trade, taxes etc.) (Government/Private); Commercial data (private, for ex. CMIE generates useful datasets). He then proceeded towards stating the challenges for external researchers like:No firm level data is released (Revenue and Commerce Ministry have an agreement to not release firm level data);Private sector does have access to the Customs data (Government officials are unclear how that happens. This data is not released by them, although it is computerized);Statistics Act says unit level data cannot disclose hence compile at commodity level; Government does internal analysis of this Customs data; and Spatial analysis possible when disaggregate to the town/village level, etc. He explained that there is an evident need for dialogue between researchers and Government. Many researchers would like to come up with a list of datasets they would need the Government to release to study some specific questions. The Government should also ask for external research, external studies to be implemented and complement the internal analysis, this would increase the analysis level of the data collected.The denomination of the datasets can also be misleading, census and survey do not design the same type of data, and some data sources do also mix both methodologies like the Annual Survey of Industries. He also pointed out some general challengescommon to all datasets:Combining data from different sources; Linking record over time; Misreporting (e.g. to avoid filling out some sort of form for those with 10 workers, firms will remain at9 workers); Ownership structure (not know which one has existed for a while; and the quality of data and misclassification. On the latter, many concepts, units of study or practices would benefit to be clarified by the data collectors and producers and other would be worth receiving specific attention. For instance it’s been observed that the distinction between plant and firm is often not clear or well taken into account by researchers. Prof. Marc Muendler and Kala Krishna mentioned that special attentionshould be granted to small firms, as many studies could be conducted and lead to key lessons. It’s therefore important to find an efficient and fixed way to collect data rigorously without putting too much burden on these firms with limited resources. Solutions were found in other countries. They put forthan example of how a rotating panel could be a solution, firms could be on it and sampled occasionally for some years. They stressed it being a real issue as it has been shown that firms with expansion capabilities decide voluntarily to remain with 9 employees not to have to comply with the extra requirements coming with belonging to the +10 PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 employees’ category. Also young firms (usually small) grow the fastest, therefore it is important to study firm size. They also mentioned the case of employees based abroad and the way they are taking into account in surveys, and the fact that sometimes firms outsource employees and therefore do not show up for the census surveys, hence there is reason to question the use of the number of workers to be a parameter for cut off.They added that discrepancies can be observed over the years in one dataset without explanations to be given. For the Government, this is the consequence of firms changing category and then disappear from the survey as they no longer fulfill the selection criteria. In general there is a need to increase collaboration and dialogue between researchers and Government on data collection and research topics. A platform to keep the discussion going about the data availability, their characteristics and accordingly aggregate all the information could serve as a tool. Regarding Data on informal sector and the NSS on unorganized employees, several researchers participating in the workshop consider that as firms move in and out of formality, the sample size is too small and a panel would be required to allow links across time. It would benefit the research on the informal sector, which represent only 1/10 of the studies conducted compared to those focusing on the formal sector. It is necessary to know how things are working in the informal sector to understand how firms can grow, and move towards formal sector (etc.). It was pointed out that panel can help reach better GDP estimates. To this the Government clarified that the Economic Census is not considered as a datasets on its own but more as a frame for other data collection exercise like the NSS. The Government is collecting this data for measuring GDP estimates, and not for policy purpose. Also improvements are being made to the design of this survey: in 6th round they have 2 lists: - year of start of operation under current ownership and- for greater than 8 workers. But Government representatives did not seem convinced bythe idea of a panel for unorganized sector, not sure about the purpose it will serve, given the high firm turnover. Many issues were also discussed regarding exports data. Prof. Kala Krishna mentioned the fact that producers are not required detailing the type of products; they only provide information on total sells. They also export directly or indirectly through intermediary (but what about information asymmetry). An approach based on client could be more efficient, as it is done in Brazil. Different classifications were used, creating barriers to combination of data: for instance DGCI data (ITC classification) vs. ASI Data (uses a different classification). Prof B. N. Goldar clarified that data is collected to match the Government agenda first, which is completely understood and recognized by the researchers. However the Government also fully recognized that allowing for more research to be implemented in these sectors will be also very positive for them. The idea of reinforcing and extending an internship program for graduates at the Ministry of Finance to other ministries like MOSPI was mentioned. PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 SESSION II The session was moderated byProf. B. N. Goldar from the Institute of Economic Growth, University of Delhi. The panelists were Prof. Marc Muendler, University of California; Prof. Kala Krishna, Pennsylvania State University; and Prof. JagadeepSivadasan, Stephen M. Ross School of Business, University of Michigan. Prof. Marc Muendlermade a presentation on what India should do to improve the data status in the country. (Please see the presentation at the end of this document.) He went on to mention that It is critical to document how respondents interpreted the question while conducting a survey and how respondents think can give new interpretation of responses. If there is over time categorical change, it should be allowed for such changes in the data collected and these changes should be carefully documented. He stressed that size, number of employees, threshold issues are potential definitions for entering or exiting a dataset.Young firms – business setups are different (that size, and growth might not capture), so it is necessary to have age and ownership information to be collected.City size distribution has lessons for studying firms [Zips Law, Granularity].It is not just the big firms, but also the small firms that matter, in order to calculate GDP.Indian small firms face obstacles to growth, and stay small. It is very important to allow research on small firms and young firms. For example, poaching shows the importance of human capital. Also young firms conditional on survival grow even faster. He gave several examples of data collection in other countries For instance there is a Belgian firm study on existing firms, with a distinction between de alio, and de novo firms. In Brazil firms are identified thanks to numbers which are consistent over different databases: 14 digits in Brazil’s taxation number, first 8 digits are industry codes. Brazilian datasets also include data on who are the clients they ship to, etc. Denmark is another stated example for common identifiers, so efficient that you can track down until the managers’ level. Maybe it is not necessary (nor advisable) to go that far and the question on how closely you want to track and share information/data on people and firms is asked. According to him, it is important to consider that formal firms can have informal employees, and informal firms can partially declare their employees to public institutions like social security systems. So the definition and distinction between formal and informal firms is subjective and remain debated. Prof. Kala Krishna talked about the sectors in Indian economy which would benefit for having more data available and more research studies implemented. She introduced by saying that a few decades ago India was the Golden Standard among emerging countries in terms of data collection and availability, thanks to a good infrastructure put in place very early on.Data collection exercises in India are very big in size for some and mobilize lots of resource mobilized. For some subjects, smaller data collection could be put in place, and better targeted. This PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 would allow answering these questions within the resource constraints of the Government. According to her the subjects that still need to be explored, but the characteristics of the data make them hard to explore until now are:Labour contracts, labour laws and sub-contracts, rules at national and subnational levels,What are the barriers to exports?; What are the most important sector for the Indian exports?; Lack of information on sellers and for buyers to know who is a reliable seller etc. She then went on to say that the form in which the exports data are available today, does not allow identifying key sectors for the Indian export sector. Disaggregated data, at the firm level would be necessary. They could be collected through small surveys done randomly, or small extra questionnaire being conducted during the conduct of a bigger survey to limit the extra costs. The resource mobilization would be minimal but, following the facilitation of research being implemented on these questions, the policy impact could be very important.How well is India serving its population, Health, Education sectors etc., remains to be explored fully by researchers, who are lacking necessary data. For Prof. JagadeeshSivadasan, firms are a tough area to study even in the US, as the question of how to define a ‘firm’ always remain.(Please see the presentation at the end of this document.)He stated that it is critical for researchers to have access to micro-level data. And that the Indian authorities would need to allow researchers accessing this confidential micro-data. He went forward to say that Examples exist worldwide of such systems, even in India, and NSSO, CSO, etc. need to consider them and create such possibility.For instance the Indian census is setting up 18 research centers (based on a recent conversation with Deepak Rastogi).Alternatives to or examples for NSSO could be:Maryland – NCAER – Panel HH surveys (this is being done); and CMIE 120,000 HH panel [if you pay approx. Rs. 7 per extra question they could add a few more for specific researchers wanting certain info] He added that many subjects remain outside researchers focus areas due to the lack of necessary data. Labor laws are a very complex with the 45 existing Central Acts, 16 Associated rules, etc. There are many layers of regulations and their impact is not studied at all due to the lack of data. The same situation exists for the study of the taxes sector. For many of the questions researchers want to study within these sectors, data collected thanks to new sections on a small sample within a bigger survey already in place could be sufficient. He went on explaining that infrastructure is another topic of high interest for researchers, as is construction and immigration and in both cases relevant data is missing to conduct rigorous research studies.He closed by saying that new problematic have emerged in traditional sectors like education and health. Nowadays corporate groups are involved in delivering services in these sectors and no data is available to study these evolutions. Prof. B.N. Goldar presented the latest efforts and initiative from the public authorities.According to him Improvements are being made within the data collection exercises arealready in place.A new initiative put in place this year allows introducing Company Identification Numbers for the corporate side of the Annual Survey of Industries.In 2012-13 ASI design has changed (previous at state level, but now at district level, sampling not at district level).Changes in the design could allow panel data to be generated; it’s currently being tested, as doubts remain regarding the design. PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 He mentioned certain key elements: NSS data covers those not in Factory Act (even greater than 10 or 100 workers); ASI covers the universe of those in the Factory Act; ASI will link to Prowess (only listed companies); MCA-ASI data can be linked; MCA2 – 1 databases are now available, covers even 100% foreign owned firms; Employer-employee tracking could be done via EPF numbers. The public authorities attending the workshop responded that they understand the interest for researchers (and not only) to be able to link ASI to trade data DTCI (daily trade returns) and to have access to more disaggregated trade data, for instance to study global chains, but the confidentiality remains an issue to be sorted out.The questions important to the Government can be answered with the data currently collected. However suggestions of changes in the collection or availability of data are taken into account by the Government. For instance with ASI, a committee reviews this type of requests. Also the need for an Annual Survey of Services was common to the Government and the researchers and a pilot is currently being run. Responding to a question from the audience, they pointed out that there is nothing in the Indian legal system that specifies that all data collection exercises have to be performed by the Ministry of Statistics MOSPI, hence it would be possible for other ministries to implement their own surveys. There is still a strong resource constraint highlighted by the public authorities. However researchers expressed their wish to be involved in this data collection effort and would be willing to help improve the datasets, their collection and dissemination. PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 SESSION III The moderator of this session was Dr TCA Anant, Chief Statistician of India and Secretary Ministry of Statistics and Programme Implementation. The panelists were Shri G.C. Manna, Additional Director General, Economic Statistics Division, Ministry of Statistics and Programme Implementation; and ShriSanjibBasu, Director, Directorate General of Commercial Intelligence and Statistics Shri G.C Mannastated that the panel can be difficult to generate, with difficulties from the fact that there is no unique multiplier for example.Some data users approach the computer centres to generate panel data on request.So it can be possible but there are also confidentiality issues, as many requests include name and address of the firms are not possible to share. He spoke about ASI having a new design at district level to respond to the requests of data users wanting district level estimates.So panel data should be possible to generate, without any risk of confidentiality issue, as there are sufficient number of units/factories, per strata per district (into 4 digit).ASI frame is also available online and it could help answer some of the researchers’ questions.ASI will introduce Corporate Index Number to allow linking among datasets.ASI is conducted at factory/firm level.A Directory of Establishments with 10+ workers is to be created from the 6 Economic Census exercises, including name, address, industry, no. of workers, type of ownership, etc.Ideally a Unique Identifier number could be given to each establishment, and this identifier would be used in all surveys. He went ahead and spoke about the Census sector: changes in design made number of factories surveyed doubled: from around 20000 to around 40000, so the design is very important, and has impact on resources and on the quality of data collected. The credibility of estimates is at stake.Regarding subcontracting, in ASI, in the employment block, there is the information if factories are directly engaged or are subcontractors. This type of information is available there.For some states all factories are surveyed every year, so for those states at least a panel could be easily generated.It is worth asking the Computer center if it is possible. He added that it is possible that data for some sectors like education, health, is to improve. But it does not have anything to do with ASI. ASI is not useful to collect information on education, health etc. as information is collected at the factory/farm level, not at employee level, so no information on education level of employees is collected. But in NSS, the employment component captures the type of ownership where respondent works, (public/private, formal/informal). While the demographic block captures some info on employees: ex educational background. His concluding remarks focused on the question of including more variables on ASI. The Government is opened to that; they are having seminars/workshops to discuss this, including with researchers. Some suggestions are discussed in articles published in the Journal of Industrial Statistics.They just had a big 2 days seminar with NSSO officers from the whole country and are currently preparing in the centre in Kolkata a list of suggestions, to be discussed during the standing committee and see how they can incorporate them. Regarding the process of inquiry, discussions take time, and researchers have to PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 understand the timeframe/schedule involved.The Government recognizes that sometimes the extraction of information is not a simple process (ex. Format cvs, method of payment) and they will be taking into account. It should be simple to improve. ShriSanjibBasu went to discuss about DGCIS. DGIS is under the Ministry of Commerce and customs data are collected by customs which are under the Ministry of Finance. DGCIS is using customs data as secondary data. They do a bit of cleaning data, especially regarding ITC code which is sometimes wrongly coded by exporters, converting into standard unit when it’s not done and then publish the report. He stated that firm level data cannot be published as per the Statistical Act. Any change on this would require a change in the Act. But it is known that it is on high demand from data users.DGCIS has improved on the dissemination aspect with the creation of an online portal to disseminate authorized data. Trade data are available for download against payment of fees by credit/debit card online. No firm level/transaction level data but country, port, commodity, for the last 15 years. So panel data is available. Region-wide data is also available.Statistical Act also forbids firm level data to be shared with Parliament, and can identify major players in export/import. India is exporting more to other developing countries in the area, quality issues are being raised while exporting to developed countries like Japan. He then confirmed that there are issues with identifiers. Exporters are given a specific code to be able to export. Some firms also have other reporting numbers like Business Identification numbers, but this one is not given by trade ministry. CIN number that is being put in place can be allocated to firms which should allow to link databases. Coding should be made mandatory by law at all stage, otherwise code disappear on documents.Also one exporter can export different products and a unique number for an exporter will not allow knowing the composition of its exports by products. For this you have to go by the classification, ITC classification. Product classification also applied with ASI. Prof. TCA Anantopened by saying that there are many difficulties to get panel data, with ASI first. Panel data play a very important role in research, but it is equally important to consider that ASI is not designed to generate a panel data or to facilitate research in industrial statistics but to get an aggregate estimate of industrial activities, specifically value added but also other key parameters. Its main purpose is to allow you to do this for the whole country every year. It’s how ASI has started. Difficulties in design were already mentioned to get accurate estimate. According to him an important element should also be kept in mind with ASI - that ideally it should be self-reporting (in theory in Statistical Act). It’s the case in many countries. In India given that the corporate governance exists, self-reporting would lead to very small or very incomplete data. So in practice ASI was put in place. He further stated that NSS design is based on physical visits of agents to firms. Worry exists that with the expected growth of the manufacture sector, resources will not be able to keep up with this system in place. So we need a much stronger self-reporting governance framework. Collection in Statistical Act PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 would permit that. Reporting structure is being developed, which can be done very easily through an online mechanism, based on corporate practices already in place.Self-reporting issue does not belong in the statistics area but in the policy and governance sector, much larger challenge to change.With a panel you are supposed to be able to identify what has been included and why and follow them up accordingly. ASI not being designed to generate panel data, no guaranty that firms surveyed remain the same over the years or that their disappearance will be documented/explained. Sample sector is based on a sample. There was even a self-inclusion criterion with the census sector, in which you can get included if you cross a certain threshold of workers. He specified that the design is not to collect panel information. Until recently no particular attention was dedicated to consistent unit numbering. It’s true that computer centre can try to generate panel data but they find it very difficult. A number of years back for which you have information on a firm varies. Changes are due to change in the design, incoherence, etc. It’s a work in progress and each version of this dataset is getting better, but it’s still a challenge, but projecting from a sample to a larger group will always remain a challenge. Now with ASI instructions have been given to do the unit numbering more carefully from one year to the next, initiate in 2010-11. But the dynamic nature of the sample sector will remain an issue.Panel are an accidental consequence, not the main purpose. He went on to add that the old vs. new design is a problem for researchers. Changes were made as push was made towards a framework to end up with a comprehensive description of the manufacturing sector or at least a much larger coverage than what was being done with NSS resources. Changes should facilitate the construction of state and central sample in a manner that permits integration. The portal for self-reporting is also part of this effort. State participation is still a challenge.Regarding requests/suggestions to add information in the ASI, it’s true that they are considered and accommodated when possible. An advice for data users would be to keep in mind the distinction between ASI data and other firm level data. ASI data is establishment data, based on record maintained at the producing establishment, from record keeping obligations from the Factories Act, including some characteristics of the workers, of the work being done, etc.ASI designs to gather those information and generate estimates. Prof. TCA Anant added that many requests received are for information at the manufacture/enterprise level, but all this information is not collected from the same person. So it creates a challenge for data collector to gather this extra information from other sources than for the data he needs, until efficient self-reporti in place. We also have to consider when the responses will be reliable or not. When a person shares information about him, they are completely reliable, but when he has to report information on other workers, other family members, information becomes less reliable, this is especially true for certain information/question (sector of activity, reliable; number of people employed, and approximation (contract employees for instance)). It is Important for researchers to put themselves in the mind of the respondents. Common numbering issue, how to link trade data and ASI data. The problem is that the exporter registration process is a very different legal process than the establishment registration under the Factories Act. Idea would be to create a unique identification system for all sorts of business, regardless of their legal status (firm, factory, exporter, etc.). The recognition of the problem PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 and this idea of a solution is there but it’s a long road until something changes,structural issue with trade data also. Trade data is very detailed level of information at the shipping bill. But still it remains impossible to identify the Indian state of origin. Producers, factories, port, all in different locations. To solve this issue would require a common reporting obligation of activities of not just the manufacturers, but also traders, etc., entities outside the manufacturing sectors,which is not available at the moment. Pr Marc Muendler then highlighted why panel could be useful for the Government as well. Standard error on GDP is big worldwide. In India it comes from the fact that the multipliers change overtime. Panel data would give information on need to change the multipliers or not. Also on the quality of data, on way to make sure that your cross-sectional exercise is correct every year is to look at panel data. (Errors from inclusion/exclusion criteria and sampling frame).Hence take the panel perspective to get more accurate results, and panel will inform on data quality. Looking at both panel and cross sectional data would give information on the gaps on both sides.It’s very interesting for Indian Government and policies, they would have the primary benefit, and secondly it will be beneficial for researchers.Government representatives still have doubt about the ability of panel data generation to help with the multiplier issue. Also Prof. MarcMuendler discussed taxes data which are not available in India by law restrictions. Taxes/tariffs info cannot be released to the public by law, this is the case in many countries not only India. It helps ensuring for honest reporting. But quite often countries also have another law that says that the statistical agency gets to ask the taxes authorities to share some information. It could be worth thinking about what kind of exports data could be played in to as statistical agency database at the individually identified firm level and then be shared respecting all confidentiality rules to researchers. That’s how most countries proceed. Prof. TC Anant responded on taxes data issues. It is actually the way DGIS proceeds to get access to taxes information. But they don’t release them at unit level data, but they have access to it. Finding a way to release the unit level data respecting all confidentiality rules and laws (both the Collection Statistics Act and the Tax Act) is currently being studied. Many researchers around the table insisted on the fact that that type of retro-connection is of no interest to them. On the opposite, a unique identifier across datasets making data completely anonymous would be a key improvement in the way data is collected and disseminate in terms of the research this would make possible. Prof. Marc Muendler mentioned then the issue about contractual workers. What the Government has in mind in this case goes actually beyond what is done by many countries in terms of including them in the employment data collection. Many countries just consider employees on the pay-roll. Maybe it is aiming to high, as getting an accurate figure of the permanent employees will already be a big step. But that type of information, if collected, would set India away from other countries and would be pretty unique. The possibility of creating synthetic panel for firms was briefly discussed. It appears that a similar exercise was conducted for household consumption data but the one attempt by a known for firms with ASI did not lead to any publication. PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 DrSandipMitra from ISI- Kolkata Sampling and Official Statistics Unit asked if ASI survey and the other surveys already conducted could include the GPS coordinates of the firms.It seems that it would generate confidentiality issue, as GPS data can be assimilated to a postal address and could permit the identification of the firm. Pincode are collected already but cannot be released now for confidentially issues. DrSandipMitra also asked if we could add in ASI Questionnaire question about ‘How many regulators do you have to face. It is an important issue in India. First idea would be that big firms are confronted with large regulators, small firms confront with state level regulators.The response from the Government is that it would be hard to find somebody/ relevant source to answer this in a firm. Operational managers? Policy makers of the firm? Regulators/firms interactions may also not happen when the records are kept. ASI is a record base exercise and this is not a question part of the mandatory recording requirements. Another enquiry was about the sampling methodology from Dr Mitra was about the sampling unit of the surveys: if employee data are not fully accurate, what could be another method of sampling, maybe taxes submitted, size of investment or else? TCA Anant responded that the Governemnt has been considering other sampling frame but has not found better one yet. Taxes could be a good idea but would need some more reflections to overcome difficulties, like the exemptions, etc. It is admitted that this topic would need further reflections has no other better sampling method come to mind. Prof. Sivadasan added that the US Business registry generates data allowing sampling including employees’ data. But this is not possible at this stage in India, among other obstacles, it would not cover unregistered/informal sector, and it requires a yearly census and many resources. Prof.Marc Muendler mentioned that it would be necessary to coordinate the efforts and changes made by Government entities among them to have consistence through different databases. The resources being limited, there is no need to have a duplication of efforts and a multiplication of methods and models.On this topic there is a consensus that the creation of a data centre could be a good initiative, if the framework put in place is respecting all confidentiality rules. On top of ASI, Economic Census data, etc., other public institutions could release data through this centre. Also it could be partially funded by researchers’ resources gathered for their studies. Prof Kala Krishna and several researchers finally highlighted that changes appearing in surveys from one year to another, like firms disappearing, etc. would need to be carefully documented and disclosed to researchers using these datasets. The Government representatives declared that this should be possible, having CSO comparing sample Yo and Y-1 of a survey, as the design is not supposed to change. A deeper reflection on how to disseminate this information as efficiently as possible is indeed necessary. PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 SESSION IV – Presentation of the Data Link Initiative Sharon Buteau, the executive Director of IFMR, presented the Data Link Initiative led by IFMR LEAD in collaboration with researchers from the Sampling and Official Statistics Unit (SOSU) of Indian Statistical Institute (ISI), Kolkata aims to bridge this gap. She went on to detail that this project intends to make an interactive web-based portal comprising of several online and offline data related activities aimed at increasing interaction between data producers, policy makers and researchers and promoting best practices in data collection and evidence-based policy. Among the central activities IFMR LEAD and its partners will aim to produce a comprehensive data repository, comprising of both primary and secondary data, accessible in the public domain. Moreover, IFMR LEAD and its partners will document the methodologies, strengths and weaknesses, following the DDI (explain DDI here) Metadata Standards. The Data Link Initiative will act as an online forum for discussion about the data and important topics of evidence based policy research and will identify data gaps in development research in India. This project is intended to be actively promoted to engage stakeholders, via digital media, but also outreach through seminars, workshops, and a comprehensive summary status of the data available to address the knowledge gaps on important development issues. The key development issues are expected to emerge from the analysis of data collected so far by various government and non-government institutions across the country. Over a period of time, IFMR LEAD and its partners believe that Data Link Initiative will become an extremely valuable resource for development researchers working on topics relevant to India across the globe and will also inform policy debates on development issues in India as well as improve data collection activities. It will also allow data producers to be more involved in discussions with those who actually use the data, providing useful feedback or improving data collection. During the initial phase of this initiative, the primary focus will be on India however in the longer term, we do envision the inclusion of other South Asians countries. She explained that this project is focusing on specific areas. The project aims to cover the data in several socio-economic areas such as: Financial Access, Industry/Firms/SMEs, Environment, Health, Education, Livelihoods, Agriculture and Macro, Industry/Firms/SMEs. The list can be modified as we further discuss the scope of DLI with the relevant stakeholders. IFMR LEAD and its partners also plan to have a panel of academic experts to oversee and guide us on each thematic area. IFMR LEAD household survey datasets will be added onto the portal, following the procedures and international standards. We intend to improve the dissemination of the primary data we collected over the years. In addition to adding the datasets, the strengths and weaknesses of the data will be emphasized as well as suggest other potential interesting studies that can be done using the data. Mrs Buteau then detailed that the main targeted beneficiaries of this initiative will be policy-makers, academic researchers, development planners, PhD students, statisticians including official statisticians etc. IFMR LEAD will actively promote this data and knowledge portal in appropriate academic and development forums through workshops, active campaigning and advertisements and via their extensive network in the development field. PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA March 27, 2015 The Data Link Initiative is envisaged as an interactive web-portal to encourage dialogue between data users and data producers and to promote better use and understanding of data from various sources. She described the overall aims of the initiative being the following: (1)Promote a culture of evidencebased policy making; (2)Increase awareness about availability of development data in the context of developing countries like India;(3)Promote a culture for critical appraisal of available datasets; (4)Create a vibrant on-line/offline forum for interaction between data producers and data users;(5)Identify data gaps and limitations of existing datasets; (6)Promote best practices in collection and dissemination of data; (7)Identify potential research questions on important development issues and the data necessary to work on them; (8)Improve the outreach of open-data innovations; (9)Contribute to the formulation of evidence-based policy thanks to the research conducted, data collected, and made available. Sharon Buteau then asked the participants about the immediate next steps for this initiative. It was advised by the workshop’s participants to start as soon as possible to publish data on the website. It will be important to have an international reach to get information on best practices worldwide. Many researchers added that it would be crucial to disseminate the data in a friendly format, with key metadata and consolidated information and for free, something that is not done at the moment. They added that it is very important that international research grants funders make mandatory the release of the data collected thanks to their funding. Things are moving in this direction at the moment. To publish this data, there will be a need for a platform or portal and the Data Link Initiative could play this role. PROCEEDINGS OF THE WORKSHOP ON FIRMS, INDUSTRY AND SME DATA IN INDIA PRESENTATIONS MADE IN TECHNICAL SESSIONS Session I:Dr Sam Asher Session II: Dr Marc Muendler DrJagadeepSivadasan Session IV: Mrs Sharon Buteau March 27, 2015
© Copyright 2026 Paperzz