tinent, and I might say that my written statement has some corrections on it and I will supply a corrected copy to the reporter when I have an opportunity to have it retyped.

I want to thank the subcommittee for the opportunity to testify and also commend the subcommittee for exploring this ground which is at once new and very fertile for planting seeds of programs to help solve the problems of achieving maximum possible employment and eliminating "structural" unemployment in our society.

I might state also that as a member of the Ways and Means Committee in the House where we have jurisdiction over unemployment insurance, this has been a subject that I have been trying to grapple with as best I can for many years, and also my work on the Joint Economic Committee ties directly in here.

I appear here in the rather unique position of having cosponsored in the House two bills which are variations on the same theme, a theme parallel to and complementary with your study of "employer encouragement of on-the-job training."

On August 24 I introduced in the House H.R. 10628, which had been introduced in the Senate by Senators Javits and Hartke. On September 9 I introduced with 45 other House Members H.R. 10934, a bill also introduced in the Senate on the same day by Senator Prouty, entitled the "Human Investment Act of 1965." Both bills use tax credits as a means of encouraging private firms to undertake the training of the manpower they need to fill their skill requirements. Of course, both these bills were referred to the Ways and Means Committee.

The fundamental purpose of both these bills is to remove the bar of taxation from the development of human resources. These bills are compatible they both approach the problem of structural unemployment by removing impediments to the rapid development of the Nation's true wealth, the skills and knowledge of individuals.

The Javits-Hartke-Curtis and the Prouty-Curtis bills differ in several respects which I will identify briefly. Javits-Hartke-Curtis would make available the approximately $3.1 billion in unused credit which was provided by the investment credit provisions of the 1962 Revenue Act to employers for investment in manpower training

programs.

This approach is limited to programs to meet specific needs: (1) to train those whose skills are necessary for national defense; (2) to retrain those whose skills have been made obsolete by automation; (3) and to retrain workers dislocated by defense shutdowns.

The investment credit allowed would be 7 percent of the allowable training expenses, with a maximum of $25,000 plus 23 percent of the taxpayer's tax liability in excess of $25,000.

The Prouty-Curtis approach would provide a 7-percent credit in addition to the 7-percent credit provided by the investment credit in the 1962 Revenue Act. The amount of the credit is the same as that provided in the Javits-Curtis bill, and the same as the credit provided in the 1962 Revenue Act for capital investments. It is in addition to the deductibility of the sum spent on training or a regular business expense. The Prouty-Curtis approach specifies six allowable employee training expenses:

(1) Wages and salaries of employees who are apprentices in registered programs;

(2) Wages and salaries of employees enrolled in on-the-job training projects under section 204 of the Manpower Development and Training Act of 1962;

(3) Wages and salaries of employees participating in cooperative education programs involving alternate periods of work and study; (4) Tuition and course fees paid by the employer to colleges or business or trade schools for the training of employees and prospective employees;

(5) Home study course fees paid by the employer to colleges or acredited correspondence schools for the training of employees and prospective employees; and

(6) Expenses to the taxpayer of organized group instruction, including classroom instruction of employees and prospective employees, and including instructors' salaries, books, equipment, et cetera, but not the salaries of trainees.

This, in outline, is the essence of the two bills I have joined in sponsoring in the House. I want now to relate these two legislative approaches to my fundamental views about the role of human resource development in the economy and its statistical measurement, and to my proposal for further changes in our tax laws to allow the free, full development of our most valuable resource--people.

The Subcommittee on Economic Statistics of the Joint Economic Committee this year began a study on the statistics measuring "wealth" in the United States. Public hearings were held on a recent study conducted by the wealth inventory planning study of the George Washington University.

I sought to drive home the point that the physical and fiscal wealth of the society is only a part of its wealth. I argued that, as a society advances economically, its mere physical and fiscal wealth becomes the minor part of its true wealth. The greater part of the society's wealth lies in the knowledge within the society, in the skills and developed talents of its people. This is an extremely difficult form of wealth to measure, yet we should make a determined attempt to measure it, least we disregard or underestimate it.

During the Joint Economic Committee hearings I expressed concern lest our present economic statistics studies be misconstrued and taken out of context to be used as the basis for national policy. For example, I pointed out that what I thought to be a basic error in the investment credit provision in the 1962 Revenue Act was the failure to realize what real wealth, true capital investment, actually is. This 1962 tax provision permitted companies to obtain a tax credit for investment they plowed back into physical capital assets.

The President's Council of Economic Advisers, being of that school of economics sometimes referred to as macroeconomists, those who deal in aggregate economic figures vis-a-vis microeconomists who emphasize the composition of the aggregates in their support of the investment credit provisions, made the same error primarily because they overemphasized the significance of gross national product in measuring economic growth and in setting Government fiscal, monetary, and employment policies.

56-057-66- -5

Gross national product, although probably the most important single statistic we have, measures only economic activity. But even so, it incompletely measures economic growth. For example, the physical and fiscal economic base must be sufficiently sizable to produce a certain amount of economic activity.

So a society must have greater physical and fiscal wealth to produce a $650 billion annual GNP than one which produces only $250 billion. However, war, major economic mistakes, or other circumstances can bring about increased economic activity temporarily which may well show up in increased annual GNP for a period of several years, just as the aftermath of a war, or of major economic mistakes can result in decreased economic activity which shows up in decreased annual GNP for a period of years.

Much of that investment and development which builds a truly strong and lasting economic base does not show up in GNP figures, particularly in the years when the base is being built. Certainly "living it up," as it were, rather than sound growth which adds to the economic base, can be deceptively present in the relatively high annual GNP figures it frequently produces.

I placed in the Congressional Record of June 4, 1963, pages A35513554, an excerpt from a book by Dr. Fritz Machlup, professor of economics and international peace, Princeton University, entitled "The Production and Distribution of Knowledge in the United States," published by the Princeton University Press in 1962, in which he set forth a better way of measuring economic growth than our traditional GNP measurement.

I would like to supply that for the record here, Mr. Chairman. (The material referred to follows:)

KNOWLEDGE PRODUCTION GROWS FASTER THAN GNP-A SIGN OF REAL ECONOMIC GROWTH

Extension of remarks of Hon. Thomas B. Curtis, of Missouri, in the House of Representatives, Tuesday, June 4, 1963

Mr. CURTIS. Mr. Speaker, one of the most common assertions made in any discussion of the U.S. economy is that our growth rate has slowed down substantially in recent years and that extraordinary measures are called for to stimulate a faster rate of economic growth. The assumption that our growth rate has slowed down is based upon an analysis of the increase in the Nation's output as measured by the gross national product computations.

As I have frequently pointed out, the gross national product does not measure real growth, but rather the level of economic activity. We know, for example, that gross national product increases dramatically during a war period, but no one would assert that this is real economic growth.

The gross national product measures only imperfectly, if at all, many of the components of real growth, such as increased leisure time, new and improved goods and services on the market, more and better education, improved health and longevity, and so forth. If we could accurately measure some of these real growth factors, I am confident that a picture would emerge of a dynamic and rapidly growing economy.

Happily, an important start has been made in making these measurements. One of the Nation's most distinguished economists, Prof. Fritz Machlup, of Princeton University, has undertaken the pioneering task of trying to assess the growth in some 30 knowledge-producing activities involving research and development, communications media, information machines, information services, and education, including that in the home, on the job, in the churches, and in the Armed Forces. Professor Machlup estimates that in 1958, total knowledge production was almost 29 percent of the adjusted gross national product.

Compared to an average annual rate of increase in gross national productat current prices from 1954 to 1958 of 5.1 percent and from 1947 to 1958 of 5.9 percent, Professor Machlup estimates that the weighted average of the annual rate of increase in knowledge production was 8.8 percent from 1954 to 1958 and 10.6 percent from 1947 to 1958.

The differences in the rates of increase, Professor Machlup points out, appear even more impressive if knowledge production is compared with everything else included in gross national product. He says:

"If knowledge production, the sector comprising 28.7 percent of total GNP, increased by 8.8 percent (or 10.6 percent over the longer period) per year, an increase of total GNP by 5.1 percent (or 5.9 percent) implies that the production of other goods and services increased by only 3.7 percent (or 4.1 percent over the longer period)."

These impressive gains in knowledge production are not fully reflected in the gross national product figures, which are commonly used as the measure of our Nation's economic growth. As Professor Machlup says, an economy such as ours, with a large service-producing sector, may not be able to show off with as large a physical growth rate as an economy which concentrates on the production of tangible goods. He points out:

"Increases in productivity in the performance of intangible services cannot be measured; indeed, most of them are in the form of improvements of quality, defying all attempts at quantification. No matter how real, how substantial, how important they are, they need not be reflected in any increased money values of input-their only measure. As we have seen in the discussion of research and development, an increase in the efficiency and productivity of these activities may eventually result in increased productivity in industries producing physical goods, but the production of knowledge does not exhibit an increase on that score. This failure of growth indexes to reflect improved efficiency in the production of intangible services has several implications. One of these relates to structural differences between economies: an economy with a large service-producing sector may not be able to show off with as large physical growth rate as an economy that concentrates on the production of tangible goods, the increase of which is shown in the index of physical production and in GNP in constant dollars."

Another important service rendered by Professor Machlup is to draw a careful distinction between growth rates, which he says can be determined only by looking back over a span of 10 to 15 years, and rates of increase over a short period of time. Almost all economists, statisticians, publicists, and politicians mistakenly speak of growth rates with reference to a period of 3 or 4 years. Professor Machlup emphasizes that an increase over 3 or 4 years may turn out to be only a short-lived upsurge, followed by a downturn around a trend line, showing no or only slow growth. Likewise, a decline may be a short-term movement around a sharply increasing trend line. This misuse of terms, he says, is especially serious when the growth policies of a national economy are debated.

Under unanimous consent I include an excerpt from Professor Machlup's book, "The Production and Distribution of Knowledge in the United States," published by the Princeton University Press in 1962, in the record at his point. The excerpt which I am including covers Professor Machlup's discussion of the points which I have highlighted in these brief remarks:

"THE GROWTH RATES OF KNOWLEDGE INDUSTRIES

"After our speculative discussion of the possible and probable interrelations between rates of increase in certain branches of knowledge production and rates of increase in total national product we are even more curious about the actually observed rates of increase. This presupposes, of course, that the poor statistical data-to no small degree based on heroic estimates or guesses-can be regarded as empirical 'observations.' But we shall proceed as if the data reflected what actually happened, even if we must make many mental reservations concerning their reliability.

"NOT EVERY INCREASE CONSTITUTES GROWTH

"In talking about the increases in the estimated values of production in recent years we shall be guilty of not avoiding the semantic sin committed by almost all economists, statisticians, publicists, and politicians of our day;

namely, of speaking of 'growth' rates although we know only rates of 'increase' over a short interval of time. Only looking back over a span of 10 to 15 years would we be justified in speaking of growth. An increase over 3 or 4 years may turn out to be only a short-lived upsurge followed by a downturn around a trend line showing no, or only very slow growth, or it may even prove to have been a temporary recovery in a longrun decline.

"This misuse of terms, serious when the growth record, the growth objectives, and growth policies of a national economy are debated, is relatively harmless in our present survey; hence, we need not be so pure in our language. We shall bear in mind, of course, that a rate of increase observed over a short period does not warrant an assumption that this increase will be 'sustained' over a longer period. When we fall into the bad habit of speaking of 'growth' instead of an an 'increase' over a short period, this does not mean that we have forgoten our own warning.

"INCREASED MONEY VALUES WITHOUT PRICE DEFLATORS

"Another snare to guard against is the identification of increased dollar sales or increased dollar expenditures with increased production, without considering the price changes that have occurred. In order to measure increases or decreases in production, we must 'deflate' dollar sales by changes in the prices of the products or, where there is no physical product and production is measured only by inputs, we must 'deflate' dollar expenditures for input by the changes in the prices of the inputs used. To do this would be a formidable task. The work involved in finding the price deflators for each industry or branch of knowledge production would be enormous, and the results would not be worth the effort. Another way of interpreting changes in dollar figures must be found.

"A crude but simple method is available: We shall not attempt to measure physical volumes of production, but shall simply use dollar values at current prices and compare the rates of change in dollar expenditures in the different branches of knowledge production with one another, with the rate of change in dollar expenditures for the total of all knowledge production, and with the rate of change of GNP at current prices.

"For the latter comparison we shall not bother to adjust GNP for changes in the status of certain products as 'final' or 'intermediate.' After all, what matters at this juncture is not the size of GNP in any one year but the rate of change of GNP over certain periods.' For this purpose the official estimates at current prices will be used; we choose the period of 1954-58 for the rate of increase over the most recent period,' and 1947-58 for the rate over a somewhat longer period. GNP at current prices was $234,289 million in 1947, $363,112 million in 1954, $442,224 million in 1958. This involves a 'growth' of 21.8 percent over the 4-year period 1954-58, and a 'growth' of 88.8 percent over the 11-year period, 1947-58. Translated into annual rates, the shortperiod increase was 5.1 percent and the longer period increase was 5.9 percent per year. These 'growth rates' were the combined reflections of physical increases and price increases. The only use to be made of these annual rates of increase in GNP at current (rising) prices is to compare with them a number of other rates of increase, observed in the same or similar periods and likewise unadjusted for price changes.

"GROWTH IN EXPENDITURES FOR KNOWLEDGE PRODUCTION

"In table IX-2 we undertake to show the rates of increase in expenditures for production in the various knowledge industries or branches of knowledgeproduction, as far as we have obtained data. The table is designed to present (1) the 1958 (or closest available) value of each item, as taken from table IX-1, to be used later as weight for the 'growth rate' in order to arrive at average growth rates for knowledge-production as a whole; (2) growth rates for the most recent period, preferably 1954-58; and (3) growth rates for a longer period, preferably 1947-58. Unfortunately, in several instances different periods have to be used because data for the chosen years are not available; and for some items we must resign ourselves to showing blank spaces.