Census of India 2001

Mobilising Support for India’s Census-

Constraints and Challenges

J K Banthia

Registrar General and Census Commissioner, India

Pretoria, South Africa

26th November 2001

INTRODUCTION

Few countries in the developing world have the distinction of having an unbroken series of a population data set coming from population censuses as India has. Beginning at least from the non-synchronous series in 1860s and ending in1871, the first census is usually dated 1872, India carried out its successive, uninterrupted population census, the fourteenth, in February – March 2001. Apart from the obstacles in planning and executing the several steps of census, much worldwide hype had been building around the pregnant expectation of whether India’s population is going to touch the magical number of a billion or not? Population projections had suggested that possibly the billion

mark would be crossed sometime in May 2000. The arrival of the billionth Indian had been timed not only to the month and date but also to the minute and seconds and made world news headlines. Against such a media hype and built up, Census 2001 was to be conducted a few months later, when in a way the result had already been declared. Suppose the preliminary results of census declared on 26th March 2001 had announced India’s provisional population to be less than a billion, all hell would have been let loose. However this was not impossible! An undercount of about 3-4 percent, not an improbable scenario in several censuses in developing nations, was all that was required for such a fiasco. The stakes involved were exceptionally high and much reputation and credibility at risk.

Census in India is carried out under the provisions of the Census of India Act, 1948. There are several articles in the Constitution, which make it mandatory for the government to undertake census. Census is carried out in two phases- Houselisting and House Numbering Operations for making the census frame for the second phase and to collect certain data on the housing conditions. The Population Enumeration phase which was carried out throughout the country between 9th to 28th February 2001 with a revisional round between 1st to 5th of March. In Jammu and Kashmir and certain snow bound inaccessible areas the population enumeration was carried out during September to December 2000. Census is carried out by the canvassing method, which involves visiting each and every household and collecting information at their doorsteps, by an enumerator. Every person within the boundaries of the country is enumerated irrespective of sex, caste, religion, nationality and age. This includes the houseless and pavement dwellers also. Census information collected is confidential by law and is used only for statistical purposes in aggregated form at the pre-designated administrative levels. It is mandatory for every household to provide information in respect of each member as truthfully as possible to the approved questions put by the census enumerator.

Although it is not possible to list all the constraints and challenges faced in mobilising support while undertaking the preparations for Census of India 2001 an attempt is made to summarize them under five main heads as follows:

1 Financial

2 Legal

3 Socio-political

4 Technical

5 Administrative

These are discussed at some length in the succeeding pages and will provide some insight of the sort of problems which the country as big and diverse as India is faced with while conducting its census. It also allows us to know how these problems were overcome and the census completed throughout the country after a gap of thirty years.

1 FINANCIAL- CONSTRAINTS AND CHALLENGES

One of the most critical inputs in organizing census is to get the requisite financial commitment and timely sanction from the government to allow the organization to plan out its activities as per the census calendar. With already a very large population, which continues to increase rather alarmingly, the country was somewhat disadvantagedly placed as it means obtaining a sanction for a bigger budget for census even if everything else remained constant. It is normal procedure in the government to marginally increase the budgetary support over and above what was spent previously. But since censuses are held only once in ten years, the finance ministry looked in utter disbelief when the census budget for 2001 was projected at Rs. 1200 crores- approximately four times over the 1991 estimated expenditure. There were two main objections- why is it necessary to conduct census in two stages and with a much smaller army of the enumerators and supervisors? The reasoning was that expenditure on account of the data collection could be reduced considerably if the above logic was accepted.

It was explained that in absence of a good frame of addresses and absence of house- numbering system in most parts of the country, it was essential that the Houselisting Operations are carried out not only to prepare a frame of enumeration areas but also to collect data on the housing condition in the country. If houselisting phase was not carried out , the undercount of the population at the country and state level could be as high as 5 to 10 percent. In such circumstances no purpose would be served of carrying out the population census and if challenged legally the entire exercise would be indefensible and untenable. Further the main advantage of producing reliable small area statistics would have to abandoned jeopardizing the entire planning process of local areas. To get better value for money spent it was also proposed by the census organization to enhance the scope of data collection during the Houselisting Operations by including several new questions. The focus of enquiry therefore was shifted from ‘WHERE PEOPLE LIVE’ to ‘ HOW PEOPLE LIVE’. The government easily accepted this shift in the purpose and objective of conducting Houselisting Operation to assess the quality of life of the people. The battle had been won but the war still remained.

The question of reducing the number of enumerators and particularly the supervisory staff and their honorarium remained unresolved. The past practice was to appoint one enumerator for 120-150 households in rural areas or for 100-120 households in urban areas. This was a time tested work allocation to enumerators, which had worked well in the past and it was decided to follow the same principle of work allocation for the present census also. For every five enumerator one supervisor was appointed to keep a watch on the progress of the work since census is a time bound operation and to provide technical support whenever problems arose in data collection during the field. The finance ministry argued for enhancing the workload of the enumerators for covering 200 and 150 households in rural and urban areas. It was also suggested to reduce the number of supervisor by providing one supervisor for every ten enumerators. Both these measures would help in reducing the requirement of enumerators and supervisors from two million to less than one and a half million and thereby reducing the expenditure considerably. However this was not found feasible on the ground that since the scope of enquiry was being increased both for the phases of census there was no scope whatsoever of increasing the workload of the enumerator in these new circumstances. Further the pilot survey while pretesting the census schedules had amply demonstrated that the enumerators work load had considerably enhanced. Increasing the number of households therefore would be virtually inviting a disaster in terms of the data quality. Given the short period of three weeks it was found to be impractical for one supervisor to oversee the work of ten enumerators in rural areas where settlements are scattered far and wide. Maintaining a constant liaison in bigger metropolises where the enumerators would be from several government offices would be extremely difficult. It was argued that insistence of appointment of one supervisor over ten enumerators would ultimately lead to virtually no control on monitoring the fieldwork of such a mammoth exercise. Both the coverage and the content of the census data would be adversely affected, hence neither of the proposal of the finance ministry were found to be practical and in view of the high risk involved were ultimately abandoned.

During the British period census enumerators were paid hardly any honorarium as it was considered a part of the essential duty of the village officers to assist government in census taking. Since 1951, after independence, a small honorarium had been introduced. For example in 1961 and 1971 about one dollar was paid as honorarium to enumerators. There was a lot of zest, fervour and enthusiasm in taking part in a national endeavour such as census. Particularly the teachers to offer free volunteer services for census considered it a matter of prestige and honour. However, things have changed considerably in recent years and now it was getting difficult to get enough enumerators even on payment basis. The proposed hike in the quantum of honorarium for the enumeration staff was from about 15 US$ in 1991 to 50 US$ in 2001. This was largely justified in view of the inflation and also the steep increase in the salaries of the government employees, which had taken place since the 1991 census. The average current salary of a teacher was about 250 US$ per month. The workload of both the census phases put together was estimated at equivalent of about twenty-one working days of eight hour each. It was argued that the census enumerator was being only partly compensated taking into account hat census work is a national effort and that the enumerator continues to get his salary as usual. The reluctance of the teachers to take up additional work except census and elections was also a factor to be considered. Further many teachers in urban areas were engaged in taking private tutions which was an extremely lucrative proposition when compared with the tough and exhausting census work of visiting house to house at odd and unearthly hours. There had been scathing criticism of the government after the 1991 census of for being penny wise and pound foolish in paying a pittance to enumerators while preparing such a huge and vital demographic and socio-economic database of the country. The honorarium paid to enumerators elsewhere in the sub-continent was also compared with the present proposal. Ultimately it was pointed out that the per capita cost on account of data collection was going to be only 12 cents and this was a very small cost for obtaining such an important database. After much discussion and in view of the compelling reasons put forward, the proposal to increase the honorarium from 10 US$ to 50 US$ was accepted much to the relief of the census organization. It may not be out of place to mention that this investment appears to be already paying as the coverage and quality of some of the provisional results of census seem to indicate.

Traditionally locating quality professional trainers for training the enumerators and supervisors has been a bottleneck. The sub-district officer who is also the local and ultimately responsible census officer is expected to function as the master trainer but in practice due to multiple day to administrative responsibilities find little time or inclination to undertake and fulfill this important role. This job is ultimately carried out either by the local census clerk or one of the experienced census supervisors who has worked in the earlier censuses or by an official of the census directorate. As a result the quality of the training leaves much to be desired. To overcome this, it was decided to appoint special selected and intensively trained Master Trainers who would take 9 training class each and also subsequently monitor the census help centers at the sub district office. Since there was no separate budgetary allocation for Master Trainers they were appointed as supervisors within the overall strength of the enumerating agency by making suitable changes. The quality of the training was reported to have improved considerably as a result of this newly adopted measure.

Some of the important weakness of the census data was in overcoming the bias against the women which somehow or the other crept in during data collection. Since part of the problem could be attended to if special emphasis and attention be drawn of the enumerator to this gap. Although the census questionnaire had been modified to make it more gender sensitive and the instruction manual also altered considerably to meet this requirement, the question was how to incorporate it as part of training. The gender biases were particularly noticeable in certain north Indian states. The 1991 census data on sex ratio, women literacy and work participation rate was analysed to assess the districts having perceptible gender biases. About 250 such districts were identified and a special capsule on gender issues in census particularly highlighting the local plight was prepared. Retired experienced senior census officials were located and given a special orientation class on gender issues in census and the methodology proposed to be implemented during training of enumerators. These officials were designated as Census Advisors and assigned the specific role of explaining to the enumerators how to make specific queries on gender issues with the help of local examples and earlier census data of the area. There was no budgetary support existed to undertake this desirable activity. Fortunately certain UN agencies such as UNFPA, UNIFEM and UNICEF came forward to support this novel approach proposed by the census organization. As a result over 200 Census Advisors were appointed in the identified critical districts to undertake special training class of the enumerators on sensitizing them in overcoming the gender biases the respondent may have while collection information from the household. It is heartening to note that the vastly improved overall sex ratio of the population inspite of a decline in child sex ratio suggest that that the enumeration of the women in the Census 2001 has possibly improved. Similarly there are enough indications that work participation rate of seem to have improved in many of the north Indian states.

Certain additional cost on account of equipment such as LCD projectors and laptop computers were also met through the assistance of the UN agencies mentioned above. These novel approaches to training of enumerators helped us considerably in making the training classes extremely absorbing and keeping alive the interest until the very last class. Certain other additional costs on account of improving the quality of paper and its printing required for census schedules were met from the overall budgetary support. In the past censuses high quality of paper for census schedules was not an essential requirement but due to adoption of the new scanning and ICR technology use high quality paper became necessary. Although the best quality of paper could not be used due to budgetary and other constraints, yet certain efforts were made to use better quality paper for census schedules than used in the earlier censuses.

The other major budgetary constraint was on account of the funds required for census data processing. Roughly half of the budgetary support was required for this activity if the census data processing of the earlier census was to be continued. In 1991 census about 45000 temporary employees were recruited in 165 offices across the country to undertake the manual data processing work for producing small area statistics and certain tables on religion, language and the scheduled castes and the scheduled tribes which required tabulation on a full count. The other tables were produced on a sampling basis and through data entry on computers. If the same methodology was to be adopted it was expected that this would require recruitment of 65000 part time employees for two years and involved opening of temporary 195 offices of about 10000 square feet area each. This proposal would entail an expenditure of about 125 million US$ in addition to cost to be incurred on the data entry of the sampled population. Since the adoption of the appropriate new technology for census data processing was yet to be examined and finalised, the budgetary proposal was made on this count was made as per the methodology adopted in the 1991 census- manual and part computer assisted data entry.

All the above constraints led to an initial estimate of about 260US$. This amounted to about 25 cents per capita direct cost on census data collection and processing. Part of the low per capita cost is the huge economies of scale, which allows certain census expenditure to be spread over and bring down cost considerably.

2 LEGAL – CHALLENGES AND CONSTRAINTS

The authenticity and the credibility of the census data, the fact that it is statutory exercise having constitutional backing and affecting the political representation of certain groups and influencing in a very large way allocation of funds from the government for several development programmes makes the census exercise extremely vulnerable against judicial intervention. Particular segments of the population think of legitimizing their claims through census and therefore look for opportunities of intervening through the census. This is because of the extreme care with which the census exercise is undertaken and its results put in public domain even for very small administrative units and particular social characteristics. Fortunately there have been very few judicial interventions, which have questioned either the process of census taking or the census data. Yet of late there have been increasing incidents wherein certain groups have attempted judicial intervention threatening at times the commencement of the census process.