annual rise in average labor quality, each with appropriate weights gives the annual rise of 2.2 to 2.4 percent in output per man-hour. I is important to understand that the rise in this most widely used meas ure of productivity reflects all three developments-improved eff ciency, investment in tangible capital, and investment in labor quality A figure of 2.2 to 2.4 percent may seem small, but it is not. Suc an annual rate means an increase of about 25 percent in a decade, an a doubling in 30 years. Few countries have long-term records tha look as good as this. Indeed, many countries would be very happy t have their current level of output per man-hour equal to what we ad in an average decade. Over the larger part of the globe output pe man-hour is less than 25 percent of the U.S. level.

The long-term rates we have been looking at are averages whic oalance out any year-to-year fluctuations there may be in the rate of productivity increase. In fact, productivity did grow at an uneve rate, as the graph on page 356 shows.

The rate of growth of productivity has undergone a number of shifts, of which one of the most interesting is the distinct change in trend that appeared sometime after World War I. Productivity has increased, on the average, more rapidly after World War I than be fore. This change in trend is visible not only in the indexes for the private domestic economy, but it can be found also in the figures for the whole economy, including Government. All available statistic make it clear that the rate of growth in productivity witnessed by the present generation has been substantially higher than the rate expe rienced in the quarter century before World War I.

There have been cyclical fluctuations in productivity also-that is both before and after World War I, productivity fluctuated with the state of business. Year-to-year rises in productivity have been greater than the long-term rate when business was generally expanding and less (often falling) when business was generally contracting.

Beyond the cyclical fluctuations, there are other changes that show up when one looks at the year-to-year statistics. These include occa sional spurts and slowdowns in productivity growth that extend over a period of years.

Of particular interest is the trend over the 15-year period since World War II. Output per man-hour rose after the war at an average rate that was distinctly high, though not unprecedently so, for a pe riod of 15 years. The postwar rate of increase, over 3 percent per year, was much higher than the rate over the full 75-year period, and significantly higher than the rate over the period between 1919 and 1960. Unusually rapid increase in tangible capital per man-hour helps to explain this spurt in output per man-hour.

But while it has been rapid by our historical standards, the postwar rate of increase in output per man-hour seems to have been substan tially exceeded by corresponding rates in some Western Europear countries, in Russia and in Japan. Their high rates probably reflect recovery from the disastrously low levels of the war and immediate postwar periods, and high rates of investment in plant and equipment-higher even than in this country. But there are undoubtedly other factors, also, which economists still must ferret out.

Year-to-year changes in productivity were appreciably influenced not only by business cycles but also by random factors. Some of these irregular changes reflect the inadequacies of the figures themselves. Productivity change is measured by the ratio of two indexes, output and input, each subject to error; and even slight errors in these will sometimes combine to produce considerable error in the ratio, just as they will sometimes cancel out one another. We cannot be sure whether the change between any particular pair of years is the result simply of statistical error. On the other hand, that the errors are not overwhelming is suggested by the fairly systematic business cycle behavior that we have noticed. We know, also, that some of the irregularities reflect not statistical error, but the impact of weather, strikes and the other real random factors to which life is subject.

Because the rate of increase in productivity has been far from uniform, the user of productivity figures must know the period to which they relate. Rates of productivity increase derived from one period will differ, sometimes considerably, from those derived from a longer, or shorter, or altogether different period. The same caution may be noted with regard to extrapolations of past trends into the future. These, the record suggests, will always be rather risky.

Thus far the rate of growth in the entire economy's productivity is the prime fact with which we have been concerned. The facts on productivity in individual industries also deserve attention, however, because they help us to understand the process by which national productivity has been raised.

These additional facts show that rise in productivity has been a general industrial phenomenon. Although there was great variation among them in the average rate of rise, every one of the individual industries included in the chart experienced an upward trend in output per man-hour, and this was just as universally true of output per mit of labor and capital. (Information on output per unit of labor is. lacking for individual industries.)

It will be noted that the statistics in the graph relate to a limited number of industries. The 20 industries which are included in the graph make up less than half the entire economy, measured either by putput or input. These industries, some narrowly and some broadly lefined, are largely from the commodity-producing sectors of the

economy. Lack of statistics prevents giving similar information for the service industries, construction, trade, and government, and ever some individual manufacturing, mining and utility industries. How ever, it is very likely that productivity has generally increased als in these.

What we know of technological developments and the other imme diate causes of productivity change in the service industries, for ex ample, supports the impression of a general rise. We know, too, that the factors that make for increasing efficiency in the use of resource are felt everywhere in the economy. Virtually all industries use me chanical power and have reaped some advantages from broadened na tional markets. No industry has been free of the drives that improve efficiency. Further, connections between industries provide channel along which new, improved or lower cost materials, fuel, power, serv ices and equipment, as well as ideas, flow to improve efficiency.

This leads us to an important conclusion about the sources of a industry's productivity gain. It is true that what happens in an in dustry is influenced by the diligence, enterprise and ability of its work ers, management and investors. But it is also true that what happens in an industry is influenced also by the quality and quantity of what the industry obtains from other industries, domestic and foreign.

Before we leave the indexes for individual industries, another com ment is desirable. These indexes are often less reliable than the in dexes for the economy at large. It should be recalled, also, that for many individual industries productivity indexes are entirely lacking a lack that makes itself felt when policies are proposed that require such individual indexes.

Finally, a word of caution is needed even about the national indexes These, like all estimates, are surrounded by margins of error. There is reason to believe that, for general historical and analytical purpose the indexes are not too wide of the mark. But this is not to say that they are also good enough to be used in directing the course of wage and prices, as some people have proposed.

The main facts we have learned may be summarized as follows: National efficiency, as measured by output per unit of labor and capital, has increased over the past three-quarters of a century at ar average annual rate of 1.5 to 1.7 percent.

Output per man-hour has grown 2.2 to 2.4 percent per year-ever more rapidly than output per unit of labor and capital combined This is so because output per man-hour reflects not only the increas in efficiency but, in addition, the increase in tangible and intangible capital per man-hour, which also has been substantial.

Rise in productivity has been experienced by virtually every in dustry, but the rate of rise has varied greatly among industries.

The trend of national productivity, whether measured by outpu per unit of labor and capital, output per unit of labor or output pe man-hour, has been significantly steeper since World War I than in prior decades.

During the 15-year period after World War II, output per man hour has risen at an exceptionally high average annual rate, mainly because tangible capital per man-hour also has risen more rapidly thai in most earlier periods.

How these facts on productivity change are related to the changes that have occurred in wages and in the prices of individual commodities, and the bearing of these facts on how wages and prices should change these questions belong in other chapters of the story of productivity in the United States.

[Statement before the Senate Subcommittee on Employment and Manpower, Sept. 26, 1963] PRODUCTIVITY AND TECHNOLOGICAL TRENDS IN THE PRIVATE ECONOMY, 1947-62

(By Leon Greenberg, Assistant Commissioner for Productivity and Technological Developments, U.S. Department of Labor)

In response to the request of the subcommittee, this paper presents some statistics of trends in productivity in the U.S. economy and some related data on technological change. The main points to be covered include: (1) Recent trends in productivity and how they compare with previous trends; (2) relationship of change in productivity to change in output; (3) analysis of trends for major sectors of the economy; (4) an illustrative statement of current technological developments.

In attempting to analyze and evaluate the trend of productivity, it is almost impossible to avoid controversy. Even those analysts who are trying to formulate objective judgments may come to different conclusions using the same body of information. In addition, productivity is caught up in the net of collective bargaining, where workers are concerned with income and job security and employers with costs and profits.

There can be reasons for differences in interpretation and evaluation. Various ratios of output to input can be computed and there is no single productivity index which answers any and all questions. Indexes may cover the economy, major sectors, or individual industries. Different time periods may be covered and may be affected by such factors as the business cycle and war production efforts. Even the word "productivity" is in some dispute, where some invariably define it as output per man-hour while others maintain that it must include capital as well as labor inputs.

There are, unfortunately, some statistical problems in measurement because the data are not always as precise as we would like them to be.

Despite these difficulties it is possible to develop a logical framework for analysis in terms of both industrial and time period coverage. When evaluating productivity for technological impact on the labor force on employment and unemployment-a useful ratio is output per worker or per man-hour. This ratio helps in studying progress in manpower utilization and its corollary, labor requirements.

It is helpful to first examine the productivity of the entire private economy. This total reflects the changes that have been going on among all the component industries as well as the shifts among these industries or sectors. For example, a shift of manpower from farms

1 General government is usually excluded because there is, as yet, no way of measuring the output of the public sector except as reflected by wages and salaries.

to factories can raise the productivity of the economy because the dollar value of output per man-hour; i.e., the level of productivity, in manufacturing is higher than in agriculture.

So the measure for the total private economy is useful as an overall indicator of productivity change; that is, of the total change in output per person or per man-hour resulting from technology and other factors affecting the structure of the economy. But it is also necessary to examine the components of this measure in order to explain and understand some of the changes which have taken place. Moreover, the changes should be pinpointed, insofar as is feasible, to the particular industrial sectors in which they are occurring. Is the impact of technological change uniformly pervasive throughout the economy, is it randomly scattered, or is it concentrated in a few key spots?

Finally, it is important to examine recent trends, but a good deal of the evaluation must grow out of some understanding of cyclical or other "wave" influences and out of comparison with previous time periods.

OUTPUT PER EMPLOYEE TOTAL PRIVATE ECONOMY

In tracing the progress of productivity and the impact of technological change, it is usually preferable to work with the ratio output per man-hour. However, trends in employment and unemployment are most frequently described in terms of number of persons in the labor force. Therefore, as a starting point, it may be useful to indicate what has happened to output per employee (including the selfemployed) based on labor force estimates.

Output per employed person in the private economy increased at an annual rate of 2.5 percent in the last 5 years (1957-62). This is slightly lower than the 2.7-percent rate of increase for the postwar period as a whole (1947-62) (table I and chart 1).

Output rose more than output per employee in 11 of the 15 postwar years and, of course, employment rose in those years. Employment declined when output per person rose faster than output. For the postwar period as a whole and for the last 5 years, on the average, output increased more than output per person, and employment rose about 0.7 percent a year.

The output per employee figures are affected both by average hours per worker and by output per man-hour. The latter is a more useful measure for studying the impact of technological change and the rest of the discussion will be in these terms. For this purpose, manhour information from the establishment surveys made by the Bureau of Labor Statistics will be used.2

OUTPUT PER MAN-HOUR-TOTAL PRIVATE ECONOMY

Productivity (output per man-hour) in the private economy increased about 4 percent in 1962. This was a relatively large increase. However, large increases also occurred in other years. Small increases of 2 percent or less occurred in some years of the postwar period. Thus, productivity changes have not been uniformly dis

2 These man-hours are based on employment and payroll data submitted by a large sample of employers. They are considered more reliable than man-hours from the labor force series. In addition, industry detail is available from the establishment series but not from the labor force series.