There was marked instability associated with black adult females. No conclusion at all about nonwhite adult males.

I have copies of the study in New York. I did not bring it with me. I will be delighted to provide them to the committee, if interested. As more recently I have begun a study and the results are not in. I can't make definitive statements about the effects of uneven industrial coverage.

I think superficially the data suggests that there probably are quite large effects. For example, in 1930, before any fair labor standards legislation, roughly-just one second-9 percent of all workers in the trade sector were teenagers. By considering I guess by 1961, 33 percent of all teenagers were in retail trade. They had not been disproportionately in retail trade until the lower coverage in this sector.

As a very early demonstration of the kind of effects that I suspect are going to show up, I have made a very simple comparison of 1930 and 1940. It is an interesting period because the 2 years straddle the introduction of fair labor standards legislation with the first minimum wage.

And on page 4 of my testimony, if you have it available, I have a summary tabulation reporting teenage shares of employment. These are people 14 to 19 years inclusive. By industry, major industry group, for the years 1930 and 1940.

That is matched against the percent of workers covered by industry with the introduction of the legislation in column 4. Column 3 is the column of interest. What happened in the aggregate is that by 1940. the teenage share of employment had fallen to 6 percent from somewhere around 10 percent.

Now other forces were operative and to say all of that is just the effect of minimum wage is uncertain.

However, what we do find is that in industries experiencing less than average coverage, their relative share of teenage employment rose; in all industries expressing more than average coverage, their relative share of teenage employment declined.

Column 3, as I have pointed out, is the column of interest. What I have done is take the share of teenage employment in the industry, in 1940, relative to 1930. Then I adjust that for what is happening at the aggregate level so that if an industry were to move just in proportion to the aggregate, that is if the teenage share were to only fall to onehalf of its 1930 level, the index would be 1.

As you see, the computed index is below 1 in all industries with above average coverage above 1 in all industries below average coverage.

That is one instance of what I would call run from cover for teenagers, when the jobs vanished in the covered sector, they simply turn to jobs that are available in the uncovered sector.

The CHAIRMAN. Let me see if I can get a clearer understanding of what you are talking about. You used a comparison of 1930 and 1940 figures.

Dr. WELCH. Yes.

The CHAIRMAN. What did you find in terms of total employment? Dr. WELCH. The percent of all workers employer that were teenagers fell from around 10 percent, there is some conflicting evidence. Some members give 12 percent, some give 10 percent. In 1930, to 5.9 percent in 1940 where we have slightly better data.

The CHAIRMAN. To what do you attribute that reduction?

Dr. WELCH. In part increased school enrollment, in part minimum wage legislation. In part a lower portion of the level of economic activity. One effect, even in the absence of minimum wage legislation is that as we enter recessionary phases, teenage employment falls relative to adult employment.

The CHAIRMAN. Now I wonder, increased school, increased schooling, economic conditions generally, is that right? And the minimum wage. Now it was in that period that new limitations were put on child labor you know.

Where does that fit into this?

Dr. WELCH. Precisely. In part it enters by the elimination of persons 10 to 13 years old from the data for 1930. And I had to make a correction.

The CHAIRMAN. I wonder-I think you missed something, doctor. Would you explain the coverage in that period, child labor, the new child labor prohibitions? The ages and the periods.

Mr. MITTELMAN. The Fair Labor Standards prohibited the employment of any youths under 18 in certain hazardous employment such as mining and manufacturing. It also required special certificates, working papers which they are called, to permit youths 14 to 18 to work in other jobs. In certain categories of industries youths under 16 and also 18 in other categories were completely prohibited. That would basically be mining, manufacturing, and I suppose many forms of transportation.

I don't know about communication. But of the first three categories listed in part A of your table 1, I would assume that some of that change was attributable to the fact that employment of youths below 18 were simply prohibited by the Child Labor Laws.

Dr. WELCH. There is no doubt in my statement I made no attempt to attribute the aggregate movement to the minimum wage legislation. I say if I may quote, it is doubtful that all of this results in minimum wage legislation but it is clear that teenage share fell most in industries most extensively covered by the legislation.

Mr. MITTELMAN. Just to amplify a previous question, the difficulty that I have in trying to determine the validity of these things is to see if you have in fact isolated all sorts of factors that could have affected teenage unemployment during this whole period.

I mentioned one, the demographic share of the population held by teenagers. We also have, especially since World War II, but it has been fairly steady beyond that, an increasing tendency of adult women to enter the work force, to the point today, that 40 percent of the work force is composed of adult women including working mothers.

How well in your own estimation, if you could explain to the committee I think it would be helpful, certainly to me, have you isolated all of the other factors that could conceivably have had a bearing on teenage unemployment?

Now I include in that at least the demographic characteristics of the teenage working force and the increase in the labor force of other groups who will be competing with teenagers for the same jobs, particularly adult women.

Dr. WELCH. In the 1930-40 comparison no attempt was made. These numbers were put together to see what kind of pattern emerged

vis-a-vis industrial coverage. I attempt no probabalistic statements. In the paper with Mark Kosters, to which I referred, we made every attempt to identify both population and labor force participation patterns by group in adjusting. And in the statistical appendix accompanying the article, a full description of the adjustment mechanism is described.

Briefly what we did is isolate the long-term growth rate in the population share of the group. And normalized on that so we could then talk about the percentage distribution in aggregate employment. Mr. MITTELMAN. Let's take between 1930 and 1940. We had a depression in the country at that time. There was terrible unemployment in many industries. Certainly some industries were hurt worse than others.

I don't have the figures at my fingertips as to which ones were hurt the worst-now we all know, whether things get bad in industry the first to be let go are marginally skilled, unnecessary people. Were things worse in mining, manufacturing, transportation, communications, and the other categories, and the other industries listed in part A as opposed to part B?

Have you isolated for that?

Dr. WELCH. Yes; there is no question. There is a very close correlation between industrial coverage and the effect of a recession on industrial employment. And I think that in part explains the legislation.

That as long as a differential impact is going to fall on an industry, if there is a wage floor, then persons whose productivity is falling below that floor become ineligible for employment. That was in fact the full purpose of the paper with Kosters. What we find is that indeed these workers are marginal and always have been marginal.

But their marginality as measured by the percent change in employment within that group associated with a 1-percent reduction of total employment was more than double by minimum wage legislation per se. Heightened coverage and heightened levels.

Mr. MITTELMAN. So you believe you have been able to isolate the effect of, let's say the general unemployment situation, on the teenage unemployment.

If you just go by the raw numbers, sure, it will look worse, this particular industry, let's say, had a 10-percent unemployment of whom more were comprised of teenagers, perhaps.

But have you, in your own judgment, been able to isolate for that kind of consideration?

Dr. WELCH. I hope so. Let me say briefly that my interest in minimum wage legislation is a derived interest. I began working on the industrial distribution of changes in aggregate employment. After isolating not only that through certain industries were certainly more affected, more vulnerable to cyclic variation.

But seeing that within those industries teenagers were becoming increasingly vulnerable, it is true they are vulnerable at any point, but after noting the marked increase in vulnerability. I saw a copy of Tom Moore's paper with coverage data. And there we knew that coverage was most concentrated and presented itself as a likely candidate for this phenomenon and hence the paper on the effects of minimum wages on the distribution of changes in aggregate employment.

Senator TAFT. Can you draw any conclusions as to whether or not many of these other factors, economic conditions being an obvious one, have any importance with regard to a youth differential?

Dr. WELCH. No question at all. The business cycle, you know, nothing could do more for teenage employment at any level of the minimum wages that we have been discussing than the same period of economic growth. That is the real story in variation and I suspect the average level of teenage employment.

Again, ordinarily entering a recessionary period the productivity of workers falls.

Senator TAFT. So you are dealing with two variables, really.

Dr. WELCH. And it is the interaction of the two that is important. Dr. MOORE. I might add that according to my statistics a drop of 1 percentage point in the unemployment rate of males 25 years and older would reduce teenage, nonwhite unemployment rates by 12 percentage points. So it is more sensitive.

Senator TAFT. Do any of you have any indication on the other side of the coin? Do you have any statistics or studies that would tend to show youth replacement of adults because of minimum use differentials in employment?

Dr. WELCH. As far as I know, there are no estimates of these cross demand effects. Just none.

Senator TAFT. Do you have any more questions?

Mr. MITTELMAN. No.

Senator TAFT. See if the chairman has any more.

Mr. MITTELMAN. I gather none of you gentlemen have had the chance to study the experience of the New York State youth differential?

Dr. WELCH. That's correct.

Senator TAFT. I would like at this point to put in the record the article by Dr. Moore entitled, "The Effect of Minimum Wage on Teenage Employment Rates;" the articles by Dr. Adie and Mr. Chapin, "Teenage Unemployment Effects of Federal Minimum Wages;" Dr. Adie's article, "The Lag in Effect of Minimum Wages on Teenage Unemployment;" and, "Teenage Unemployment and Real Federal Minimum Wages," by Dr. Adie.

[The prepared statement of Dr. Welch and documents referred to follow:]

Statement of Finis R. Welch, Graduate School, City University of New York and National Bureau of Economic Research, Before the Senate Subcommittee on Labor of the Senate Committee on Labor and Public Welfare

June 6, 1973*

Mr. Chairman and Members of the Subcommittee:

Thank you for the chance to comment on the Fair Labor Standards Amendments. I testify in behalf of a youth and old age differential under the assumption that existing minimum wage legislation will be updated by raising the nominal minimum and possibly by extending coverage to sectors not currently included. Although my comments focus largely upon teenagers, similar implications hold for persons above 65 years.

I will not explicitly consider workers, who in the absence of minimum wage legislation, would have earned more than the minimum. Whatever effects hold for these persons are purely derivative of effects on workers who would have earned less than the minimum. To the extent that low productivity workers become less desirable to firms following a legislated increase in their cost, it is clear that high productivity workers become more desirable.

For persons who would have earned less than the minimum, legislated wage floors are at best a mixed blessing. There is the obvious potential to increase earnings, but with this is the burden that these workers must find an employer who perceives their labor to be at least worth the legal minimum or be unemployed. It is always true that workers have incentives to find jobs offering the greatest satisfaction. That without legislation, some earn less than the minimum is proof in and of itself that these

96-478 O 73 11