NEED FOR A QUINQUENNIAL CENSUS

The business of government-federal, state and local-is people. To serve them most effectively and efficiently, cities must pinpoint their services to people where they are at a given time, not where they were five years ago, eight years ago or ten years ago. In their planning, cities are concerned with projecting where people of different ages, income and other characteristics-are likely to be next year and five years from now. The ability of local city governments to assess trends and make plans depends to a large extent on the availability of accurate population and housing statistics. The need for accurate local area data is growing at a rapid rate and basic governmental problems at all levels are now and will continue to be shaped by the dynamics of population change. In the years to come, local officials will need frequent and detailed socio-economic statistics in the following illustrative areas:

(a) Planning of schools and other educational institutions-intimately related to the shifts of population within the city;

(b) Planning and design of commercial facilities for community uses— related to the population, housing and income distribution:

(c) Planning and recommendation of residential areas in the city:

(d) Planning of municipal facilities and services for all areas in the city-fire, police, water, sewer systems and health and welfare services; (e) Planning of major transportation arteries, including in particular public mass transportation; and

(f) Planning evaluation and recommendation for urban renewal areas in the city. For several years after the decennial census, city officials will have reasonably reliable statistics on which to base their decisions.

However, the pace of urban life is dynamic-the American population is highly mobile, advances in technology change the types and number of jobs available at a quickening pace, the need and demand for services alters as these changes occur. Because of this mobility and the shifting characteristics of the population, much of the information gathered during the decennial census concerning small areas (such as police districts, health districts and other categorical subdivisions) becomes obsolete over a period of ten years. The most difficult problem for the cities is to know of the internal population shifts within the city and changes in the quality of housing in particular areas. As the Census Bureau has indicated, for the country as a whole, 10% of the housing stock is lost in ten years and 25% of the housing units at the end of a decade are new. In some localities these percentages vary even more.

As I am sure you are aware, many cities in this country have been losing population to the suburban fringes or bedroom communities surrounding them. Cities receive federal and state financial aid, to a large degree, on the basis of various formulas involving population statistics which usually rely on the decennial census.

It is estimated that approximately $10 billion annually are now allocated by the federal government to the states or their subdivisions according to formulas which use population or housing data as a major component. In addition, large amounts of money are allocated by the states to municipalities annually with population as a major component in determining allocations to any one local governmental entity.

Although relatively accurate estimates can be made during the intercensal period for states and large metropolitan counties, estimates cannot be prepared with any reasonable degree of accuracy for the smaller areas. As officials in these areas are forced to use data that are more than several years old, they can be seriously misled.

Estimating techniques may be effective for several years beyond the census date, but have proven unreliable over longer periods. For example, one of the factors that appears to be inadequately analyzed by many cities (and perhaps contributed to the numerous complaints that followed both the 1960 and 1970 Census) was typical household size. In the last ten years, the average number of persons per household dropped from 3.33 in 1960 to 3.17 in 1970 for the country as a whole. I am certain the decline was even greater than the national average in some cities. Apparently the number of occupied housing units had increased in a number of cities but were occupied by typically smaller households. This decrease in typical household size came to light only upon analysis of the census

data. As a consequence sample surveys or special censuses in selected areas are not an efficient means of providing the basis for developing needed current estimates.

SCOPE OF A QUINQUENNIAL CENSUS

A five year census should be as comprehensive as a 10 year census. Any reduction in the content of the census would reduce or might even nullify the utility of the census for its users. Similarly, the substitution of a sampling procedure for complete enumeration in the mid-decade census would deny to local agencies and to other users critically needed information regarding small areas. General trend data give little in the way of indications that can be useful in assessing in which small area or areas these changes may have occurred, or to what degree. For city purposes, citywide figures or metropolitan area wide figures (which could not be broken down reliably by the Census Bureau to less than estimated changes for cities or SMSAs) would have only marginal value. To be valuable, the data must be comparable to those collected in the decennial census so that true changes in conditions can be ascertained and measured.

Planning for a wide range of local activities, including urban renewal, housing, highways, utilities and other public services, requires statistics for areas as small as city blocks and census tracts. Data on age, sex and race are basic for computing rates by which public health and welfare programs in local areas are planned and administered.

Increasingly, the duties and powers of governing units, the eligibility for certain benefits from state and federal funds (not necessarily allocated on a per capita basis), the applicability of certain laws, and even the number and salaries of public officials are tied to federal census counts.

Census data provide the base of information needed in carrying out a number of major programs involving cooperation between federal and local government agencies. These include such programs as those for manpower, poverty, low rent public housing, rent supplements, homeowner interest subsidy and housing for the elderly and handicapped persons. Urban renewal, model cities and urban planning assistance programs require up-to-date information on the number and characteristics of persons below the poverty income level, employment and unemployment, and the occupational skills of those in the community. There is need also for information on age, sex and race and the family composition of the population affected, their level of education and their recent migrations as well as population shifts within the city and between the city and its surrounding suburbs.

A replication of the decennial census is essential to insure optimum information is available for use by the cities. Opponents of a quinquennial tabulation have cited cost of such a survey prohibitive. We feel it would be a net saving for all by eliminating the necessity for undertaking local individual surveys and the need for the numerous private canvasses and local public surveys. Other surveys conducted by the Census Bureau, such as those covering agriculture, manufactures, wholesale and retail trade, etc. are all done every five years. Only the census of population, housing and employment is conducted decennially. It is our understanding that the cost of a 25% sampling is around $180 million compared to $220 million for "regular" census. The saving is not worth it.

MANDATORY RESPONSE

Statistics gathered on a voluntary basis are less likely to reflect adequately and accurately the information needed for the various purposes outlined above. We think that responses to a five year census ought to be mandatory, as they are for the ten year count.

CONCLUSION

One hundred seventy years after the establishment of a decennial census, the role of local government has broadened, the pace of events has quickened. The decennial census is an anachronism. A basic census every five years is essential to the effective, efficient and responsive functioning of all levels of government. Mr. WILSON. As our last witness for this morning we have Miss Mary K. Nenno, director of program policy and research, National Association of Housing & Redevelopment Officials.

I am sorry, Miss Nenno, you have had to be the last witness and that you had to wait all morning, but are we looking forward to your statement.

STATEMENT OF MARY K. NENNO, DIRECTOR OF PROGRAM POLICY AND RESEARCH, NATIONAL ASSOCIATION OF HOUSING AND REDEVELOPMENT OFFICIALS

Miss NENNO. Thank you, Mr. Chairman. I understand a woman always has the last word.

I am here this morning representing the National Association of Housing & Redevelopment Officials, which as you know is the association composed of the administrators of local housing, urban renewal, and housing codes enforcement agencies throughout the country. Of course, we have a great deal of interest and concern over the census information.

We are here in support of the mid-decade census, and particularly in support of the housing component in a mid-decade census.

We hear a great deal about shifts and changes in population. The fact is a great many changes are taking place in the housing supply, and we hear a great deal now about abandonment of housing in cities, which is a phenomenon that came up rather quickly; and we really did not realize the extent, the degree to which this was happening among cities in the country. If we had had a mid-decade census in 1965, I am sure we would have caught some of this earlier, and perhaps would have been able to get at some of the solutions earlier. So we are very strongly in support, naturally, of the housing component in a middecade census.

It has already been noted here that there are two areas of need as far as a mid-decade census is concerned. One we are particularly concerned with relates to the national use in the Department of Housing and Urban Department's grant assistance programs; many of these programs rely on census data as a basis for distribution of funds. I have attached to my testimony a typical example of the application which a local housing authority has to submit, and you see that it is heavily based on availability of census data.

Also, the national allocations of the Department of Housing and Urban Development funds are based on changes in housing supply, changes in conditions of housing, and things of this sort. So we see a two-level need. One is for national programing, allocation of resources; and secondly, at a local level, there is a great use on the part of local agencies in determining, for example, which areas in the city require treatment, which areas are declining. We need to know something about changes that are going on in the inventory in order to program housing assistance for low-income housing.

We need to assess the capacity of the local housing inventory to handle the needs of those who are displaced, and our local agency people use census data for all of these purposes.

I would also like to indicate that we really need data for small communities and for small areas within larger cities. I have also attached to my testimony a breakdown of the size of communities that are now participating in the urban renewal program and in public

housing. Ninety-two percent of all local housing authorities and 35 percent of all public housing units are in communities of less than 50,000. Seventy-five percent of all urban renewal agencies and 27 percent of the grants approved are in communities of less than 50,000. Now these are the communities that really require census data, both to prepare their assistance applications to the Federal Government and, secondly, for their own local programing and planning purposes. So that we really believe that if a mid-decade census is going to be helpful to them, it has to have some detail for small communities and, again, some detail within cities on a block basis in order to do the kind of planning that is required.

We are recommending something close to the level 3 census. I cannot be precise about this, because I do not know to what extent the cutbacks that we are recommending bring us between level 2 and level 3. The reason that I mention level 3 is because it is the only one that I see that does provide data for small communities and for blocks within cities. So I think we have to express our recommendation in terms of level 3 census.

However, we are proposing some cutbacks from what was proposed in 1970. At the bottom of page 3 in my statement we say, first of all, that the prime need is for a 100-percent count of both population and housing units, with a few basic characteristics of each.

This does not mean it has to be the full array of characteristics that we had in 1970, but we need some few basic characteristics of both population and housing units. We need this for two reasons. One is as a benchmark to show the changes in these total inventories, and, secondly, as an accurate base for local communities to undertake their own sample surveys.

I think this is going to be increasingly important. Many communities are trying to improve their local capacity to do surveys and, obviously, the census, as complete as it is, cannot supply all the needs for local data. But it is absolutely essential they have some kind of an accurate base on which to base the samples; and this is why we would recommend a 100-percent count in both population and housing, with a few basic characteristics of each.

Second, a 100-percent enumeration for small communities and for blocks in larger cities. I believe the current block schedule, for example, has a 24-item array, and we are suggesting that there are four items that as a minimum, would be required to get some idea of the changes that are taking place in terms of blight and decline in these neighborhoods.

Now this is less, again, than was required in the 1970 census.

But we feel that in a mid-decade census, if we had these items, plus a 100-percent count of population and dwelling units, we would be able to make some kind of estimate as to what is going on, without having the full array that we had in the decennial census.

Then, of course, we would like a 20-percent samplying for other key housing and population items on a census tract basis. It would include things like income and some of the other key pieces of information. We would like to make a particular point about the census and the survey of the structural quality of housing. If you remember, in the

1970 census we dropped for the most part the direct enumeration of housing condition. This was a tremendous loss to localities who had used this kind of data.

What we are nowing trying to develop in cooperation with the Department of HUD, is a standard format for a survey which could be undertaken by a locality so that you could combine the results from these local surveys. And if we had in the mid-decade census a sufficient direct enumeration to get structural data, using the same format, for the Nation as a whole, for regional areas, and for some standard metropolitan areas, we believe we would be able to begin to match up the local surveys and get a national picture. I think we could get better results, actually because the enumeration of structure condition is a very complex and expert job; and as we know, some of the results we got from using the census for this purpose, were not all that good.

We believe that we can do this if we get Federal assistance to develop the standard format and to enable local communities themselves to have the capacity to undertake these kinds of surveys.

Finally, in terms of the annual housing survey which has been proposed, which we strongly support, this would provide data on the components of change in the inventory for the largest standard metropolitan areas. We can see this extended, perhaps not covering every metropolitan area every year, but on a 3-year cycle, so we could begin to get this kind of information on annual basis for a few selected areas and maybe every 3 years for others. Then, the load on the middecade census to provide that kind of information would not be as great, although we would still like some general benchmark data.

I think, basically, what I am saying is that we do want a mid-decade census that provides for some detail for small communities and for blocks within large cities. We see some possibilities and opportunities for cutting back some of the detail from the 1970 census enumeration. Of course, we would be pleased to work, as we have done in the past, with the Census Bureau and others to attempt to define these items. Mr. WILSON. Thank you very much, Miss Nenno.

As you will recall, many of the problems that were raised last year prior to taking the 1970 census had to do with the housing information that was asked, particularly the one about the showers and the number of units, the number of families using a single shower, and so forth. And yet this is the very type of information for which your organization, I suppose, has great need.

Miss NENNO. It is absolutely essential.

Mr. WILSON. I wonder if the problems we had were not caused by the way in which the questions were asked. If there had been different wording of the questions-some of them were later reworded-I wonder if this is not the real answer to such problems.

Miss NENNO. I am sure that is very true. And also I think that the census in 1970 did do a much better job in their public information regarding the census, and I think it could be improved still further by showing appropriate examples of how this information is used and so

on.