188 UNEMPLOYMENT IN THE UNITED STATES

say-roll figures of nearly 1,500 establishments employing, according to the
latest information compiled by our bureau of labor statistics, 374,115 men and
women. You can see there the character of industries and mercantile firms
that furnish the figures.”
Manufacturers of metal and machinery, wood products, chemicals and oils,
textiles and foodstuffs, as well as public utilities and merchandising houses,
builders and contractors, are among the contributors that appear upon the
list he handed me. More than 70 subelassifications of industry and commerce
appear on it.
“Statistics such as these are not guesswork,” Mr. Boyd continued. ‘‘None
more trustworthy can be found, for they come direct from the employers themselves.
 Certainly, if they do not know the employment trend in their own
plants no one does. Each month they furnish the number of employees on their
pay rolls. By comparing these figures with those of previous months and years
a, pretty good idea is gained of how the particular industry stands, and to a
large extent how promising is its immediate future. These comparisons also
enable one to strike a balance, to understand which increases and decreases in
employment are directly attributable to seasonal fluctuations such as are found in
building and construction, textiles; the canning and packing undustries, and
which are due to unusual and unexpected causes. Their value and reliability
are enhanced by the fact that they show the employment trend at its source,
sometimes before, sometimes after the newspapers have carried ads for the
industry or employment offices have sent workers to it. But at the source.
nevertheless.
“It is not safe, sometimes, to make the most obviously logical deductions,
sven from accurate statistics, no matter what their source. For instance, a
report for one month may show a falling off of employees for one industry. If
you have an abiding faith in figures you'll naturally conclude the employment
trend in that field is downward and the business outlook poor, when, as a matter
of fact, the decreased pay-roll force may be entirely due to difficulty in securing
the necessary labor because of a shortage of competent help or possibly because
of a temporary dispute over wages or working conditions. The volume of
orders to be manufactured may never have been higher in the industry's history,
yet the bare figures would belie that condition. Or, from another angle, our
job-printing classification might show a marked increase in the working force for
a certain month, due solely to an emergency such as getting out election ballots
or mail-order catalogues or telephone directories, none of which would give the
true picture of the prevailing trend in the local printing industry.
“The ratio of applicants to jobs available at a public employment office is
also a substantial barometer. It is not bullet-proof, of course. For example, it
is axiomatic among employment men that during hard times workers stick to
their jobs, while the jobless, with more or less regularity, apply for work.
“But what happens when jobs become numerous? The jobless continue to
apply, or obtain work, but many of those who have had jobs quit them in order
to get better ones. They quit the potboilers for better-paying opportunities.
In other words, the number of people applying for work at our office or at any
given point does not necessarily prove that the employment situation is poor.
An increasing number of applicants for work may, in a period of great surplus of
jobs, be proof positive that a condition of general unemplovment has ceased to
exist.
“Cold figures and percentages must be mixed with a dose of experience before
we are safe in proclaiming or forecasting a trend. Employment work holds a
paradox at every turn. But these two factors—employers’ pay-roll data and the
ration of applicants to jobs available—are to my mind the best guides for finding
the trend of employment; provided, always, they are compiled by people who
have been jostled about by the perplexities and paradoxes of modern business.”
“The value to the public,” he concluded, ‘of such statistical data and interpretations
 is in direct ratio to the timeliness with which they are released for
public review.”’

CLEARING HOUSES FOR WORKERS
Which, in the light of our present-day limitless leave to print at public expense,
may goad even the most erudite to cackle: “ Ain’t it the truth?”
Employment data such as Illinois sponsors, New York, Massachusetts, Californis,
 Pennsylvania, and many other commonwealths in similar measure collect
and distribute to their citizens. Qur Federal Reserve banks, too, in periodic
bulletins and through the press contribute statistical data of similar tinge; and on
a acale still larger. the Federal Bureau of Lahor Statisties at Washington.