Page images
PDF
EPUB

the views of an observer outside the Federal Government. I am, of course, not qualified to discuss from the perspective of the Federal agencies the statistical programs in the area of wealth measurement which they have underway or plans they are making for the future, but as a user of Federal statistics I am happy to present my views on the present state of wealth measurement.

With regard to Federal statistics on the distribution of wealth, one immediately thinks of the personal wealth statistics derived by the Internal Revenue Service from estate tax returns. Historically these have given us our best information and they are widely used.

Mr. Natrella is testifying before this task force with respect to this body of data, and will, I am sure, provide an informed view of its strengths and weaknesses and the implications of recent changes in the estate tax law for future statistics on wealth distribution. It is, of course, important to recognize that this source of data, although it covers a large part of total wealth, cannot provide information on the wealth holdings of most of the people since only a small fraction have estates large enough to require filing a return. For many questions on the distributive impact of budget and economic policies, therefore, this source is grossly deficient.

But the size distribution of wealth derived from estate tax returns is only one of the sources of information available in the Federal Government on wealth and its components. While other sources do not give size distribution information, they do provide aggregative estimates on the tangible and financial asset components of wealth and their ownership by institutional sector. By putting these different estimates into a system of sector balance sheets we can obtain a better picture of the nature of wealth holding by different groups in the economy and their change over time. Once this larger picture is developed, a number of different techniques can be used to detail within such a comprehensive framework the size of wealth holdings by institutional groups and by individual households.

With respect to tangible wealth, the Bureau of Economic Analysis of the Department of Commerce has built up stock estimates based upon perpeutal inventory methods. These estimates cover not only the structures, equipment, and inventories held in the business and government sectors but also the owner-occupied housing and consumer durables held by the household sector. The BEA estimates are of course based on a wide variety of information and are directly related to the expenditure data contained in the national income accounts. In addition, benchmark data for housing comes from the Census of Housing, for farm assets from censuses of agriculture, and for plant and equipment from special surveys on plant and equipment expenditures conducted by BEA itself.

The basic information on financial assets and liabilities is integrated into the flow of funds data published by the Federal Reserve Board. The basic information for the flow of funds comes from banking statistics collected by FRB, the Statistics of Income of the Internal Revenue Service, and the quarterly financial reports collected by the Federal Trade Commission. The flow of funds accounts not only show the year to year changes in the assets and liabilities of the different section relating to consumer finances for some time, and in this connection financial assets and liabilities existing at any moment of time.

Besides the efforts of Federal agencies in collecting and compiling wealth information, research workers outside of the Government have contributed greatly to knowledge in this field. The Institute for Social Research of the University of Michigan has been providing information relating to consumer finances for some time, and in this connection I understand Prof. James Morgan has testified before this task force on this subject.

Again, Prof. James Smith of the Pennsylvania State University, who I understand has also testified, has in recent years been one of the main contributors to the measurement and analysis of the size distribution of wealth. As I am sure Professor Smith reported, he has made a major use of estate tax returns to cast new light on the distribution of wealth. Recently a new study of the distribution of wealth has been made by Prof. Stanley Lebergott of Wesleyan University. He used the statistics of income data relating to the individual income tax to generate distributions of wealth by income class. Essentially, he capitalized the property income flows received by different income groups, and then reconciled his results with known aggregate information about wealth holders.

Finally, wealth research has a long and honorable history at the National Bureau of Economic Research. Many studies were done under the sponsorship of the Conference on Income and Wealth, and many others were carried out by an illustrious progression of investigators, including Simon Kuznets, Morris Copeland, Raymond Goldsmith, Robert Lipsey and others.

Recently a specific project aimed at constructing national balance sheets and measuring the distribution of wealth was undertaken at the NBER within a National Science Foundation project on the measurement of economic and social performance. Raymond Goldsmith has been drawing on the tangible wealth data of BEA and the financial asset and liability data of the Federal Reserve Board to draw up balance sheets for several dozen sectors and subsectors of the U.S. economy for the period 1947 through 1975. This work is now nearing completion and preliminary balance sheets for the household sector are given in the attached tables 1 and 2.

[The tables referred to above may be found at the end of Mr. Ruggles' prepared statement.]

Mr. RUGGLES. Professor Goldsmith has made estimates of the balance sheets of the top 1 percent of the wealth holders as well as for all households, on the basis of the benchmark distribution provided by James Smith. As a part of the same NBER project, Prof. Edward Wolff of New York University has allocated the assets and liabilities to individual households on the basis of the microdata contained in the Census public use sample and the IRS tax model sample of individual income tax returns. Combining these various sources of data with additional information on consumer expenditures for durables, Professor Wolff has estimated the distribution of assets among different social and demographic groups and has analyzed the impact of inflation on the distribution.

As this account of the state of the measurement of wealth and its distribution makes clear, the present situation can best be described as fragmentary and uncoordinated. It might appear logical that the straightforward way to collect information on the distribution of

wealth might call for obtaining it directly from individuals and households. Certainly the classic 1962 Survey of Consumer Finances is still used as the best study of the distribution of wealth that has been made to date, and a new survey along the same lines is very much needed.

But such a survey, while important and extremely useful, would be insufficient by itself. The household sector is only one of the wealth holders in the economy. Financial intermediaries, pension funds, nonfinancial corporations, and governments all are important for any analysis of holdings of tangible assets and/or financial assets and li

abilities.

A comprehensive approach to the study of wealth is needed to show the interrelationships among the various sectors and to provide an understanding of the process of wealth accumulation and its ultimate ownership. The basic sources on which wealth measurement depends need to be better coordinated with each other, and serious gaps in basic data need to be filled.

There is no doubt but what the BEA perpetual inventory method is one of the best approaches available for keeping track of tangible wealth, but this approach needs to be further supported with better benchmarks and more basic information.

Thus, for example, the census of housing needs to be expanded to cover nonhousing structures, since these are an extremely important dimension of wealth. Better information also is needed on consumer durables. At the present time some information on consumer durables is available from the 5- and 15-percent samples of the population census, but it is to be hoped that further light will be cast on this component of wealth by the BLS consumer expenditure surveys. I understand annual consumer expenditure surveys are contemplated for the future, and important information relating to consumer durables and other components of consumer finances could be obtained from this source by instituting special inquiries from time to time. Statistics Canada has successfully carried out special surveys on consumer finances which have then been related to consumer expenditure surveys, permitting the development of information on the distribution of household income and wealth in Canada. Such data are now being provided on an annual basis.

In terms of the broader measurement of financial assets and liabilities for all sectors of the economy, the importance of the quarterly financial report collected by the Federal Trade Commission should be recognized. Unfortunately, the coverage and completeness of this report has suffered in recent years, as the FTC has wished to concentrate its efforts on its new line of business report. The SEC has shown an increasing lack of concern for providing general economic information, as have the regulatory commissions in different areas.

As a result, important financial information relating to wealth has not been available. The reduction in the statistics of income sample size has further reduced the amount of information which can be obtained. It is hoped, however, that the tabulation of the forms submitted. to IRS by exempt organizations-nonprofit institutions, pension funds, et cetera will add a new dimension to the data.

Unfortunately, there is little coordination or integration of the data originating in different Federal agencies. Most of the benefits of wealth data are external to those agencies which are responsible for producing

the basic data. The producing agencies are reluctant to assume the costs involved in constructing series which they themselves do not use, and using agencies are reluctant to pay for what has traditionally been free. No single agency has responsibility or authority for measuring wealth and its distribution, though many agencies are peripherally concerned with some aspects of wealth. In such a situation, both the quantity and quality of resources devoted to wealth statistics have suffered.

The confidentiality of data sources poses still other problems. The Census Bureau by law cannot disclose to other agencies information concerning the identity of individuals or companies reporting to it. Generally speaking, other Federal statistical agencies follow similar rules. In some cases, such as the Census public use sample or the tax model sample of individual tax returns, it has been possible to remove the identification of respondents so that individual reports can be made available. In other cases, the producing agency may for a price be willing to provide special tabulations.

In actual practice, however, such special tabulations are difficult and expensive to arrange, so that they are seldom a practical alternative. As a result, much of the available data is locked within the producing agency, and natural rivalries among the Federal statistical agencies often reinforce the effect of confidentiality restrictions in preventing the flow of information. In other words, the decentralization of the Federal statistical system substantially reduces the amount of information which is provided and prevents the building up of comprehensive measurements of wealth and its distribution.

The combined effect of decentralization and privacy restrictions not only inhibits the efficient use of existing information, but it also tends to encourage Federal agencies to duplicate one another's activities, substantially increasing costs to the Government and reporting burdens on respondents. The duplication of information collected from respondents is usually not exact, however, so that the resulting statistics produced by competing Government agencies cannot easily be reconciled with each other, and the reporting burden on respondents is increased by the need to produce overlapping but slightly different kinds of information for different agencies.

It is unfortunately true at the present time that the Federal statistical system is rudderless. There are no effective coordinating mechanisms, and the different Federal agencies are often involved in what amounts to statistical warfare. In the wealth area, it would be extremely useful to assign the responsibility for developing wealth measurement to one Government agency.

In my view, wealth measurement should be closely integrated with income measurement, and therefore the Bureau of Economic Analysis should be given the responsibility for constructing sector balance sheets which would accompany its sector income accounts and provide wealth distributions for households which could be related to household income distributions.

This task force obviously recognizes the importance of wealth data for the evaluation of the distributive impacts of budgets and economic policies. I feel certain that the other witnesses have testified better than I can with respect to the implications of wealth data for these purposes.

I have been particularly impressed by the current use of the tax model by the Office of Tax Analysis of the Treasury Department in its study of the distributive impact of changes in the tax laws and Government expenditure programs. I feel certain that introducing wealth information into household microdata sets would make it possible to analyze the impact of taxation and Government expenditure programs on wealth distribution. In terms of my own experience, I can cite two instances in which the measurement of wealth has been important in shaping my views on specific questions of economic policy.

Several years ago in designing the project on the measurement of economic and social performance for the National Bureau of Economic Research, I made a preliminary analysis of the change in household sector balance sheets between 1959 and 1969. This tabulation is shown in the attached table 3.

[The table referred to may be found at the end of Mr. Ruggles' prepared statement.]

Mr. RUGGLES. The striking thing that came out of this preliminary examination was that the entire increase in household wealth in this decade was attributable to (1) household acquisition of tangible assets-consumer durables and housing; (2) the increase in value of previously held tangible and financial assets resulting from inflation; and (3) the acquisition of pension rights. Households did no new financial saving over this decade.

Although pension rights are, of course, a form of saving, it is not a form that is controlled by households. Households generally have little influence on how the pension sector invests its funds, and in most cases the pension contribution is withheld by the employer or union and the wage earner has no control even over its amount.

In other words, during this decade households were not net investors in business. Aside from pension rights, they had no net accumulation of financial assets. These findings were further reinforced by a study of disposable saving of households over the period from 1947 through 1975, the results of which were presented to the Joint Economic Committee last November. The role of household saving in the saving and capital formation process in the economy is a matter of considerable

concern.

In the recent past, for instance, there was much discussion of the question of whether the deficit of the Federal Government was crowding out business capital expenditures, and considerable attention was focused on the sources of saving. It is my personal belief that, because of inadequate information on the distribution of wealth, the process of saving and capital formation is not well understood, and more information on this topic can help to clarify the central problems of economic policy relating to saving and capital formation.

Second, closely related to the question of household saving is the role of pensions and Government securities held by households during periods of inflation. As is reflected in the balance sheet for households in constant 1972 prices shown in table 2, the value of both Government bonds and pensions seriously eroded during the period from 1972 to 1975. These categories of assets are traditionally held by lower income

« PreviousContinue »