The success of Providencia's 1975 Industrial Census was due in large part to the care taken in advance planning, not only in the formulation of broad objectives and concepts, but in the detailed forecasting of each operational step. An important feature of this detailed planning was the development of cost estimates. One reason for their need is that without an approximate idea of what each step would cost, all the available funds could have been spent long before the work was completed. The preparation of step-by-step cost estimates, and their consolidation into larger totals representing intermediate phases of the entire project, is the process called budgeting.

1.1

Coordination of budget estimates

NSO's Budget Coordinator, who is administratively attached to the Office of the Director, is responsible for reviewing the budget estimate prepared by the NSO divisions, advising the divisions on accepted techniques for preparing budget estimates, advising the Executive Committee of problems which may have arisen in the course of their preparation, and advising all concerned individuals of progress made in their preparation. In his review, the Coordinator makes certain that the budget statements are clear and understandable, that they include all operational steps, that they delineate the areas of responsibility of each division, and that the cost estimates are realistic.

1.2 Participation of supervisors

It is NSO policy to have persons with administrative or supervisory responsibility for

carrying out an operation participate in preparing the budget estimates for that operation. This is not always possible, because of the long lead time required for the preparation and review of budget estimates. In the 1975 Industrial Census, the budget estimates were completed and approved in October 1974, but the major part of the work began in February 1976. In the meantime, some of the supervisors who had helped prepare budget estimates had either left NSO or had been transferred to other positions so that they were no longer responsible for carrying out the operations covered in their budget estimates.

2. ESTIMATING TIME AND COST

The basic form used by NSO for budgeting all projects (except small jobs which can be completed rapidly and informally) is Form 120, which is the Time and Cost Schedule. An example of this schedule, partially filled out to illustrate how it is used, is reproduced as exhibit II-4-1. In the industrial census, Forms 120 were filled out for all individual operational steps, and later consolidated into group totals and grand totals for the entire project. A useful feature of the individual Forms 120 is the number of physical units, such as the "screening interviews" shown in exhibit II-4-1. The number of physical units is not normally carried forward to the group totals or the grand total. For operational steps involving payment to outside organizations (such as printing or equipment rental) the number of man-weeks is not applicable and is omitted. the case of the county agents employed in the industrial census, the Forms 120 included estimates of their time even though they were not

In

actual performance. Some items of quantity or cost were found to have been incorrectly estimated, and the budget was revised accordingly. Where cost estimates had to be revised upward, funds could be transferred from the amount "not specified," shown in exhibit II-4-3. Eventually, however, this fund was exhausted and it was necessary to transfer some funds from other budgeted NSO projects. A carefully prepared budget should require no more than 10 to 20 percent of the total to be allocated to unexpected costs.

2.4 Overhead charge

An important addition to each time and cost schedule is the overhead charge. An example of the application of this item is shown in exhibit II-4-1. For the industrial census, the overhead cost was estimated at 15 percent of the direct cost of each operational step. This percentage covered a part of the salaries of individuals attached to the Office of the Director, of the heads of divisions, and of others who were called on for occasional work in connection with the census. The charge also included a proportionate amount of incidental expenses such as office supplies, informational services, transportation (but not the transportation of enumerators in the field, which was included in the estimates as a direct cost), communications, and other miscellaneous expenses not charged as

direct costs.

As noted in Unit I, the NSO was charged in the 1970's with responsibility for collection, processing, and publication of basic statistical data in all fields except foreign trade, agriculture, and employment. To carry out its programs in the economic fields, NSO has a permanent staff of about 40 statisticians and other professional staff such as computer specialists. In addition, there is a more-orless permanent nonprofessional staff in the central office consisting of punch-machine operators, clerks, typists, and clerical supervisors. The field staff, which was greatly expanded during the 1970 Census of Population and Housing, was later reduced to about 30 persons attached to the 12 provincial field offices.

To meet its requirements for the 1975 Industrial Census, the NSO management could draw on (a) statisticians in the Manufacturing and Mining Division, available for the industrial census because of changing work requirements; (b) statisticians to be transferred from the Trade and Services Division, available for the industrial census because of the completion of a survey of the distributive trades; (c) part of the time of the 30 persons attached to the provincial field offices, available for the data collection period by postponing other survey work such as population and housing surveys; (d) the 216 county agents whose services were made available by the Minister of Agriculture; and (e) selected non-professional personnel. Before making specific assignments, it was necessary to estimate the personnel requirements, both in total and in detail. For these estimates, an essential preparatory step was

the completion of the time and cost schedules discussed in chapter II-4.

2. BACKGROUND INFORMATION

Since the magnitude of the collection and processing workload is directly related to the number of industrial establishments, and personnel requirements are related directly to workload, the MMD planning staff had prepared estimates of the number and size of establishments to be covered in the census. The estimates were based primarily on the 1963 Industrial Census, plus whatever information could be gathered from governmental and private sources on changes since 1963. These estimates are shown below in exhibits II-5-1 through II-5-4.

2.1

Number of establishments and value added, by industry and size

The number of industrial establishments for the country as a whole is presented in exhibit II-5-1. Establishments not operated as part of households are termed "regular" and separate estimates are presented for each type. In addition, estimates of value added are presented, because they measure economic importance and provide a yardstick against which to judge the reasonableness of the costs of data collection, processing, and publication. It will be noted in exhibit II-5-1, for example, that the 24,800 "regular" establishments with less than 10 persons engaged account for nearly 16 percent of the total value added by all the "regular" establishments. Certainly this is too large a percentage to consider omitting this group of establishments from the census. However, further analysis might have revealed

that the smallest group (for example, establishments with 3 or fewer persons engaged) could have been omitted without seriously invalidating the results. Their omission could have accounted for a substantial saving in collection and processing costs. However, the MMD planning staff did not recommend omitting the smallest establishments, because there was uncertainty about the accuracy of the estimates, particularly when comparing narrowly defined size groups. To illustrate, the planning staff could have recommended omission of the smallest group because they estimated that it accounted for only 2 percent of value added, when in fact it accounted for a substantial 4 or 5 percent. By the time this discovery would have been made, it would have been too late to change the cut-off point. In view of this and similar uncertainties, it was decided to try for 100-percent coverage of the "regular" establishments in the 1975 census.

The MMD planning staff recognized in advance that collection of data from households would be a costly feature of the census, even though limited to a relatively small area sample. The decision to cover households was based largely on consideration of the magnitude of the estimated value added--121 million dollars. That this estimate was not far from the true figure was borne out by the limited amount of data on household industry gathered in the 1970 Census of Population and Housing.

The data in exhibit II-5-1 on electricity and gas plants formed the basis for another decision affecting personnel requirements--that of enumerating them entirely from the central office of NSO. The cost of training field personnel in the special features of the electricity and gas questionnaire was thereby eliminated; instead, the part-time services of a small group of experts in MMD were enlisted to survey this group. These individuals were well-trained in data collection procedures,

and they required only brief written instructions. The enumeration of these establishments, all of which were large, was conducted almost entirely by mail and telephone; less than 10 personal follow-up visits were required when the establishments had not returned their questionnaires by the due date.

2.2 Number of establishments, by province

In exhibits II-5-2 and II-5-3, the number of "regular" establishments in each province is shown. It will be seen at once that the most important province in terms of its industries is Estrella, with over a third of the establishments. For a centralized government such as Providencia, with its NSO as well as other government agencies located in the city (county) of Estrella, there were obvious advantages in having so many of the reporting units nearby. Fewer enumerators were required because distances between establishments were shorter; where establishments were widely scattered, much of the enumerator's time was taken up with traveling from place to place. In data collection, the accessibility of senior management personnel in the central office facilitated obtaining decisions on new problems; and in the planning stages of the census, pretesting of questionnaires was made easier.

Over 80 percent of the country's mining activity is concentrated in Catedral, Girardot, Tres Hermanas, and Zapata. Because of this, FOD made a special effort to secure the services of enumerators with experience in the mining field. With minor exceptions, such individuals were found among the county agents in these four provinces.

2.3 Rural and urban enumeration

Exhibit II-5-4 shows estimates of the number of rural long and short manufacturing forms, by province, compared with the total number of forms. These estimates are very important in