rl 6496 P93 09 A g073 z3 k~1_- ti$ s tf_: r E ' f; Si. pr -I m P" P~ -~ ~~~ r crr A i~~~~~hit 4._,r~-~ I;B *i; f.?:"J:: i -: -I, ~;,; a~i-;4a:; ";;~ -~~ - - - - - - - - - - -; - S-::: 77~~~~~~~~~~~~~~~~~~~~~~~:~: ~ ---;:1_:~~ I:~~I —~ ---.&~~-..~;.i —F - ~ ~ ~. _p:"~:-'4-'-';~ -;- ~ ~ ~ ~ X b~~~~~~~~~~~~~~~~~~~V r;:::::~C —.::;; -:)s-);:_:-..~~~~~~~~~~~~~~~~~~~iem:~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~":~- a ~~~. — c ~~~~-::: ~ ~~;.;,; ~~A I - ftli.: ~ ~ ~ ~ ~ i 7I I 1-~~~~~~~~~~~~~ / ~" VNWiItOM HGN G IN~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I i~~~~~~~~~~~~~~~~~~~~~~~p~ ~ ~ ~ ~ ~ ~ ~ ~~,., 411- ~ ~ ~ 1 4 ~~~:~~~~p". ~ ~ ~ 4 4f j -i-/c~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~- -~ I-~. '~i~-\ ~ c NI / I:.-I- ':-. i ~,- -, — ii: `- -- -~\I-i -e rr-: i I= —:I. ? II 'i ii 1 —,: `1,, :j,, -.L,=.- i t -i. --- v c -7::G; x. i ii C I a a THE ECONOMIC CONDITION OF THE PRINTING INDUSTRY IN NEW YORK CITY A WAGE ARBITRATION STUDY OF THE BOOK AND JOB BRANCH eL - 0- ~~~~~,. ~ -A BYpC\ DAVID R. C9RAIG, Ph.D. Assistant Professor of Industry Unicversity of Pittsburgh.~ I, New York New York Employing Printers' Association 1925 y4q6 c9 I?q Copyright, 1925, by David R. Craig \ -\ t-~Ls PREFACE Since 1907 arbitration has been the established procedure for settling wage disputes between the employers of the book and job branch of the printing industry in New York City and the printing trades unions with which they deal. Every year has brought differences of opinion, often serious differences, concerning the scale of wages which the industry should pay. Both the employers' association and the unions harbor individuals who are temperamentally belligerent, but the typical self-control of each group has sufficed to restrain them and to permit peaceable settlements. Each year, since 1919, these shrewd groups have laid down the rules for the next arbitration fight. Wages, they agree, must be determined by a consideration of at least two factorsone, the cost of living, and the other, the economic conditions of the industry. The first of these is quickly settled, for there is plenty of evidence to show whether living costs are higher or lower; especially useful are the figures published monthly by the Bureau of Labor Statistics. But the economic condition of the industry has not been so easy to describe. No systematic analysis had been made of the existing information, and the comments of the disputants on this score have more often been grounded in the shifting sands of fancy than in the more substantial stuff of fact. This essay is an attempt to supply the need for that more solid foundation. It covers so short a period, only four ears, that its conclusions cannot yet be used with complete confidence, but must first be tested in the light of events which occur in the next few years. Yet there is enough evidence to begin with, enough to warrant preliminary and tentative con-.V Vi PREFACE clusions which should help to put the discussion of economic conditions in the industry on thq same footing of objective fact which characterizes the analysis of living costs. The writer is indebted to Mr. F. A. Silcox, Director of Industrial Relations of the New York Employing Printers' Association, for having brought so interesting a problem to his attention and for having assisted in many ways in the solutions which were reached. He acknowledges with gratitude the assistance of Dr. Frederick R. Macaulay in solving some questions of statistical procedure; and the helpful criticisms of Professor Henry R. Seager and Professor Wesley C. Mitchell of Columbia University, under whose guidance this study was made as a doctoral investigation. His heaviest obligation is to Mr. C. A. Hokinson of the New York Employing Printers' Association. Mr. Hokinson's remarkable familiarity -with the printing industry and his resourcefulness in running down the data with which to answer each new question as it arose, were combined with an obliging disposition to make the task a pleasant and almost an easy one. D. R. C. University of Pittsburgh May, 1925 CONTENTS PAGC CHAPTER I. Introduction 1 CHAPTER II. The Data 8 CHAPTER III. Statistical Treatment of the Data. 22 CHAPTER IV. Fluctuations of Production, Employment, Sales, and Payroll. 28 CHAPTER V. Fluctuations of Overhead Costs. 41 CHAPTER VI. Fluctuations of Total Costs and Profits. 49 CHAPTER VII. Printingand General BusinessConditions 55 CHAPTER VIII. The Economic Condition of the Industry 68 vii I LIST OF CHARTS PAGE CHART 1. The billings of two groups of printing establishments, September 1920 to December 1922 18 CHART 2. The mechanical payrolls of two groups of printing establishments, September 1920 to December 1922...... 19 CHART 3. Press impressions, overhead costs, and total costs, 1920-1923...... 44 CHART 4. Press impressions and the ratio of overhead costs to total costs, 1920-1923... 48 CHART 5. Press impressions and profits (latter not adjusted for seasonal variation), 1920-1923. 51 CHART 6. Press impressions and profits (both adjusted for seasonal variation), 1920-1923.. 52 CHART 7. Press impressions and the sales of 15 electrotype foundries, 1920-1923.... 57 CHART 8. Press impressions and total advertising in magazines, 1920-1923...... 58 CHART 9. Employment in printing and in New York State factories, 1920-1923.... 61 CHART 10. Press impressions and the volume of manufacture, all lines combined, 1920-1923.. 62 CHART 11. Printing billings and general business conditions 1920-1923...... 64 CHART 12. Printing billings and bank debits outside New York City, 1920-1923.... 65 CHART 13. Volume of manufacture, all lines combined, press impressions, and profits, 1920-1923. 71 ix I I i I I, i I i I i I I i i CHAPTER I INTRODUCTION PURPOSE AND WARRANT The purpose of this investigation is to discover some measure of the economic condition of the book and job branch of the printing industry in New York City, for use in wage negotiations and arbitration proceedings between the Closed Shop (Printers' League) Branch of the New York Employing Printers' Association, and the various trade unions with which it deals. After the so-called "vacation period" of the late fall of 1919, which was tantamount to a strike for higher wages in the face of increased living costs, the Printers' League entered into contracts (dated October 1, 1919) in which the basis of further negotiations and adjustments was more closely specified than had ever previously been the case. It was expected by the various unions that the cost of living would continue to rise; and by the employers that a limit to the possibility of increasing wages to hitherto unheard of levels would eventually be reached in the industry itself, when wages had been raised to the point which the industry could not afford to pay. Accordingly the contracts, which were to run for two years (until September 30, 1921), provided that the agreement was to be opened for readjustment by either party on October 1, 1920, "only as to the rate of wages to be paid; such readjustments to be based upon the increased cost of living and the economic conditions of the industry at the date of readjustment."' On October 1, "'Arbitration Decisions in the Wage Adjustments, December 24, 1920," by the Closed Shop (Printers' League) Branch,- N. Y. E. P. A., p. 3. 1 \ 2 THE NEW YORK PRINTING INDUSTRY 1920, the case was opened by the request of Typographical Union No. 6, as well as some others. The demand of Typographical Union No. 6 for a horizontal increase of $18 was met by a conciliatory offer of $4 by the Printers' League. When both demand and offer were rejected the case went to arbitration. The union depended largely on cost of living figures to support its demands for an increase, and the Printers' League relied on data purporting to show that the industry was partaking of the general economic conditions of the country at the time and was therefore unable to afford an increase. This was the first attempt of the industry to state its economic conditions, and the measures which were used in the argument may be briefly summarized as follows: the level of wholesale prices; the limit of European credit (not quantitatively treated); interest rates (referred to, but not in quantitative terms); cancellation of orders and the buyers' "strike" in general business; the liquidation of surplus stocks; the number of failures and the liabilities involved; the volume of building; turnover of mail-order houses; unemployment; volume of stock and bond sales; prices of agricultural products; changes in automobile prices, especially Fords; bank clearings; abandonment of different kinds of printing (not quantitatively specified); decreased number of pages in periodicals and commercial printing; loss of advertising and advertising contracts (figures from the National Publishers' Association); the amount of printing being sent away from New York City; amount of overtime and night-work in the industry; and falling off in the return on the investment.2 The array of evidence introduced by the Printers' League on the subject of the economic conditions of the industry found the union unprepared to rebut; but after two weeks of preparation, the union introduced some further measures of economic conditions, as follows: unemployment in the industry; advertising 'This list and the one next following were taken from the shorthand report of the proceedings in the arbitration case of Typographical Union No. 6 against the Printers' Imue, October and November, 1920. INTRODUCTION 3. rates in magazines; magazine circulation; additions to plant equipment made and contemplated; profits of the Mergenthaler Linotype Company; failures in the printing industry; volume of advertising; the price of machine composition; financial statements of certain employing printers (recording capital, surplus, and percentage surplus is of capital); and number of premium men.3 Such measures of the economic conditions of the industry as these, taken at random and bearing no demonstrated relation to any obvious measure of the ability of the industry to afford an increased wage scale, were not convincing to the arbitrator. He said: "The difficulty was that not even the accredited representatives of the industry, had at their command the facts required to justify their apprehensions as to its future."4 In interpreting this decision, and the seven others that were rendered at the same time, to the membership of the Printers' League, Mr. E. A. Kendrick and Mr. Max Rosett made the following statement: "The Employing Printers when called upon to present definite and authentic facts to show the economic condition of the industry, were not in a position to do so, and if in any one particular their presentation of their cases was in a degree lacking in effectiveness, it was on this point. It therefore seems apparent to your representatives on the Board of Arbitration, that the New York Employing Printers' Association should in justice to the industry put itself in a position to furnish at any time definite, authoritative and conclusive data as to the economic conditions of the industry in New York City. The most effective and economical method of accomplishing this plan, is something for the Association to work out, but it may not be out of place to suggest here that the desired result could be achieved, through the appointment of a duly accredited firm of Certified Public "A "premium man" is paid more than the union scale for his work; the amount-of the premium being set by individual bargaining. "'Arbitration Decisions in the Wage Adjustments," etc., p. 6. 4 THE NEW YORK PRINTING INDUSTRY Accountants, to prepare annually or oftener, if necessary, a survey of the condition of the industry, based upon copies of income tax returns or duly certified financial statements to be submitted to them, for that purpose, by the members of the Association-all such reports and figures received from the individual concerns to be kept strictly confidential by these accountants and the deductions drawn therefrom to be used only in the form of a Composite Statement, as indicating the general economic conditions of the trade."5 One such accounting survey was made in the fall of 1921 by Lybrand, Ross Brothers and Montgomery, covering 48 firms. Another was made in 1922 by Klein, Hinds & Fink, in which 39 concerns were included. These surveys accomplished their immediate purpose with the current arbitrations, and have not been repeated. The accounting exhibit, when used as an arbitration device, has two limitations. Its facts refer to the industry only at a given time and do not yield a direct statement of tendencies. And useful comparisons which might be made between book and job printing and other industries are effectively precluded by the very form of the findings. The present approach was preferred in the hope that it would escape these limitations. The occasion for this inquiry, then, was found in the desirability of discovering, if possible, some "definite, authoritative, and conclusive" measure of the economic condition of the industry. METHOD It was clear at once that one limitation would be imposed on the investigation. Employers and unions in the printing industry have never reached an agreement concerning the division of the receipts of the industry between capital and labor. If such a contract existed, it would be possible to lay down an accounting procedure to be carried on jointly by 'Ibid., p. 15. INTRODUCTION 5 employers and unions, and an automatically adjustable wagescale could be evolved from the findings. But although this industry has gone farther than most industries in the direction of accommodating discussion of conflicting interests, it has never arrived at the point of saying that of all the income of the industry a certain percentage, no more or less, will go to its capital and a certain percentage to its labor. -There are no indications that this question is being pressed by either side; indeed, there are obvious reasons to suppose that neither side would welcome such an agreement. Without agreements of this sort it is inevitable that any discussion of the economic conditions of the industry should be cast in terms of fluctuations. If economic conditions means ability to pay, (judging from the origin of the phrase in the agreement of October 1, 1919, it means precisely that), and if it has never been settled just how much the industry can afford to pay in wages to its labor, then one is justified in assuming, as do both the employers and the unions, that the question is one of more or less ability to pay. The arguments in the briefs of both employers and unions are couched in these terms; the industry is said to be less able to pay than it was last year, or more able, or else it is foreseen that the industry will be less able to pay during the coming year than it has been in the past. Neither party to the annual contracts has ever formulated its own notion of anything more definite than this. And it was accordingly necessary to confine the investigation to the rise and fall of different features of the industry. The method of the inquiry, then, was to obtain as many measures as possible of fluctuation in the industry, to compare them with each other, and to compare them with measures of fluctuation in general business and general manufacturing. Apart from the main purpose of the investigation, it was hoped to learn something of the industry through comparing the fluctuations of its different elements; and in a restricted way this purpose was accomplished. 6 4THE NEW YORK PRINTING INDUSTRY SCOPE The printing industry is divided by trade parlance into several branches, of which this investigation concerns only one, the "book and job" branch. The book and job printing branch of the industry includes such services as book printing, including periodicals but not newspapers; job printing, including such materials as stationery, tags, labels, advertising which is sent direct to the consumer of the advertised goods, and the miscellaneous small jobs too numerous to classify; and commercial printing, such as mail-order catalogues, telephone books, etc. The book and job branch does not include such services as publishing of all sorts,6 newspaper printing, binding, blankbook making, engraving and die-sinking, wood engraving, steel engraving, photo-engraving, stereotyping and electrotyping.7 Further light is thrown on the work of the book and job branch by an examination of the trade unions with which the Closed Shop (Printers' League) Branch of the New York Employing Printers' Association enters into contracts and agreements. They are, Typographical Union No. 6 (hand and machine compositors,) Printing Pressmen's Union No. 51, Job Pressmen and Job Press Feeders Union No. 1, Printing Press Feeders and Assistants Union No. 23, Paper Handlers Union No. 1, Paper Cutters Union No. 119, Mailers Union No. 6, and Bindery Women's Union No. 43. But not all of the members of these unions are employed by the members of the Closed Shop (Printers' League) Branch of the New York Employing Printers' Association. This association is composed exclusively of printers in the book and job branch of the industry. The relation of the book and job branch to the entire in'But printing and publishing are not always distinguished in the reports of the Census of Manufactures. tCertain firms in the book and job branch have departments or subsidiary companies which do binding, photo-engraving, and electrotyping. INTRODUCTION 7 dustry in New York City is shown in the following table. It will be seen that the number of establishments, the average number of wage earners, and wages in the book and job branch are more than half of the total for the industry in New York City, and 35 per cent of the total in the value of its products. These figures include publishing, which is not properly a part of the book and job branch. The Census reports do not distinguish them; but the number of publishing establishments is relatively small, and the proportions indicated would probably not be altered seriously if they were neglected, or subjected to a small discount. Book publishing without printing is included in the table to show approximately the size of the discount necessary. TABLE 1 IMPORTANCE OF THE BOOK AND JOB BRANCH IN RELATION TO THE ENTIRE PRINTING INDUSTRY IN NEW YORK CITY$ Number Average of no. of Value of establish- wage- Wages products ments earners Printing and publishing, book and job......... 1,488 21,358 $38,319,163 $141,737,856 Printing and publishing, music............... 40 380 613,425 7,485,760 Printing and publishing, newspapers and periodicals.............. 758 14,485 27,838,489 252,475,386 Total.............. 2,286 36,223 66,771,077 401,699,002 Percent book and job is of total.............. 65% 59% 56% 35% Book publishing without printing'.......... 215 269 289,918 34,005,726 _~,,-, 5U. 5. Census of Manufactures (1921), "Printing and Publishing and Allied Industries," 1924, Table 33, pp. 42-44. This table does not include figures for "allied industries." OFigurs for 1919, from U. S. Census, "Fourteenth Census, Manufactures," Vol. IX, pp. 107. CHAPTER II THE DATA MEASURES OF FLUCTUATION The measures of fluctuation finally chosen were of two sorts, physical and financial. They were found in records secured by the Bureau of Industrial Relations of the New York Employing Printers' Association from the manufacturing and cost records of individual firms.' The purpose of this collection 'was originally the same as that which led to the present investigation; but no conclusive use had hitherto been made of it. Two physical measures were adopted. The first of these was a measure of production. It was recorded in terms of press impressions2 made in each month by each reporting concern on three types of presses: platen, cylinder, and rotary. Every printing concern keeps a record of the number of press impressions made in the course of its daily manufacture for the purpose of planning subsequent work. The number of press impressions varies directly with the size of editions, and beyond certain minimum limits it varies directly with the size of the job. For example, a pamphlet of 20,000 copies will show twice as many press impressions as a pamphlet of 10,000 copies, and a pamphlet of two signatures3 will show twice as many impressions as a pamphlet of only one signature. For this reason the number of press impressions makes an excellent index of the volume of physical production in the printing 'Overhead costs, from which total costs and the profit and loss items were computed, were secured especially for this study. 'One press impression is registered every time type and paper meet in a press. 'EBach signature runs through the press twice. 8 THE MEASURES 9 industry, and its value is enhanced by the ease and rapidity with which it is available to the investigator. The disadvantage is that it shows neither the expensiveness nor the surface size of the jobs which it records. For instance, an edition of 100 cheap letterheads is reported as being equivalent to 50 copies of an elaborate 32-page pamphlet containing many illustrations.4 But it is almost impossible to secure information concerning the surface-size of different jobs, especially a physical index; and the expensiveness of the work is a pecuniary, not a physical measure, by its very nature. It is necessary to explain why press impressions were used instead of the current index of the United Typothetae of America, chargeable or productive hours. The latter index might have been chosen, but it was eliminated because an hour's work in one department of printing does not compare with the same work in another, either in quantity or quality; and such a measure does not reflect changes in the rate of speed at which work is produced. Press impressions were considered more useful for the purpose at hand, for when properly weighted they are comparable at all times, they ignore the work of the composing room and the bindery, which is indispensable yet only incidental to printing (for the result of printing production is the printed sheet), and they reflect changes in the rate of production. The second non-pecuniary measure selected was the number of employees on the payroll of the reporting concerns. Not all of the number reported monthly were employed full-time over the month; their employment for any part of a month suffices to get them on the list. Nor does this measure report accurately the number of hours worked; short time and overtime are disregarded in this index. Six pecuniary measures were secured. The first of these is the billings, or gross receipts., The 'Fifty copies of such a pamphlet would of course require two printings. 'The small percentage of billings deductible on account of bad debts may be safelyg disregarded, so that gross receipts and billings are practically identical. 10 THE NEW YORK PRINTING INDUSTRY amount of work put through every month may or may not coincide with the amount of work billed. It is the trade practise to hold up billings until their corresponding jobs are completed; as a result it frequently happens that a job which is incomplete at the end of a month will not be billed until the following month. An exception is found in the December billings, when for purposes of taxation many companies typically bill the work to their customers on an accrual basis, with the tacit understanding that the bills need not be paid until the work is wholly completed. The second pecuniary measure is the mechanical payroll, consisting in the amount paid in wages to "productive" labor by the reporting concerns. The third money measure of fluctuation is the purchases. This represents a sum which is paid out month by month for materials bought by the reporting concerns, such as paper, ink, and type, but not items bought for capital account, such as machinery. The fourth is the overhead costs. These are described6 as rent and heat, light, power, insurance and taxes, depreciation, bad debts, spoiled work, department direct expense (such as oil, waste, proving paper, tympan paper, etc.) office stationery and postage, advertising, cartage and carfare, miscellaneous expense, office payroll, and sales payroll.7 The fifth is the total costs, which is the sum of purchases, mechanical, payroll, and overhead costs. The sixth is the profit and loss items. These are found by subtracting the total costs from the billings. In the main, the same group of firms contributed this information, but not all of them furnished all the information. Complete information was available for 13 companies, 14 supplied purchases, 15 billings and number of employees, 16 payroll, and 17 furnished production data. The maximum pos"Report of the American Printers' Cost Commission, by William H. Sleepeck, Chairman to -the United Typothetae of America Convention, Toronto, October, 1921. 'TSee below, p. 41, for justification of this classification of overhead costs. THE MEASURES it sible was used in each case; but in computing total costs and profit and loss items, it was of course necessary to make separate calculations for the 13 firms who submitted the whole set of figures. These figures cover the period beginning January 1920 and ending December 1923. They were reported monthly. In order properly to weight the different kinds of press impressions before aggregating them, it was necessary and possible to secure additional information on costs and speeds of operation for the three kinds mentioned. The data on which this weighting8 rests were taken in part from a cost survey9 made by the United Typothetae of America for the purpose of urging the adoption of its cost accounting system by the employing printers of New York City; and in part from records of the Cost Department of the New York Employing Printers' Association. REPRESENTATIVE CHARACTER The number of firms reporting their figures has been mentioned. Although these firms are among the largest in the book and job branch of the industry, it is necessary to inquire into their representative character before any conclusions based on so meagre a number of reports can be.accepted as trustworthy. The names of these firms are as follows: American Colortype Co.; Blanchard Press, Inc.; A. T. Delamare Co.; Inc., Eilert Printing Co.; Charles Francis Press; Federal Printing Co.; William Green, Inc.; Hamilton Press; Mail & Express Job Print; McGraw-Hill Publishing Co.; L. Middleditch Co.; C. J. O'Brien, Inc.; Periodical Press; Redfield, Kendrick, Odell; Riverside Press; Stirling Press; U. S. Printing & Lithographing Co.; and Wynkoop-Hallenbeck-Crawford. There are two ways in which the'relation of this group of $See below, pp. 23, 24. ""Survey of the Printing Industry of New York City," the United Typothetae of America, January, 1920, table 38, p. 32. 12 THE NEW YORK PRINTING INDUSTRY firms to the entire book and job branch of the printing industry in New York City may be demonstrated. The first is to compare certain aspects of the group with comparable figures for the entire branch as found in the reports of the U. S. Census of Manufactures. Table 2 below shows this comparison. As noted previously, the Census does not distinguish between printing and publishing, and the group under consideration TABLE 2 COMPARISON OF 14-17 REPORTING CONCERNS WITH THE ENTIRE BOOK AND JOB BRANCH OF THE PRINTING INDUSTRY IN NEW YORK CITY.10 No. of Av. No. of Cost Value establish- wage- Wages of of ments earners materials products No. of No. of Mechanical firms employees payroll Purchases Billings Printing and publishing, book and job.. 1,488 21,358 $38,319,163 $36,944,117 $141,737,856 Percentage of reduction'1.... 11.5 1.28 9.2 11.1 26.2 Estimated book and job printing alone. 1,319 21,085 34,793,800 32,843,320 104,602,541 Reporting concerns, 1921 totals and averages..... 14 to 17 2,321 5,608,145 2,182,864 11,393,600 Percent reporting concerns are of entire book and job printing..........94 to 1.14 11.01 16.12 6.65 10.9 r"Figures from U. S. Census of Manufactures (1921), "Printing and Publishing and Allied Industries," 1924. "Computation from figures of U. S. Census, "Fourteenth Census, Manufactures," pp. 1066-7. See below, p. 13. THE MEASURES 13 - includes no publishing.'2 In order to render the figures comparable, the figures for book publishing without printing, which are given for 1919 but not for 1921, have been ascertained, their percentage relationship to "printing and publishing, book and job" computed, and the 1921 figures for "printing and publishing, book and job" reduced by that percentage. The remainder, after this transaction, should give book and job printing figures with only a small margin of error due to an altered relationship between printing and publishing in 1921 as compared with 1919. One concludes from the foregoing table that the present investigation concerns approximately one tenth of the entire printing (book and job) industry of New York City. That the firms represented here are large is demonstrated by the fact that one hundredth of the total number of establishments handles nearly one-ninth of the value of the business. A similar comparison may be made with respect to the mechancial payroll between the concerns figuring in this investigation and the total membership of the New York Employing Printers' Association. Table 3 shows this comparison. TABLE 3 RELATION OF FIRMS IN THIS INVESTIGATION TO TOTAL MEMBERSHIP OF THE NEW YORK EMPLOYING PRINTERS' ASSOCIATION 13 Mechanical Percent payroll, represented 1921 by 16 firms 16 firms................................. $ 5,608,145............ Closed shops (Printers' League)............. 18,909,249 29.6% Closed shops (unaffiliated)................. 759,544 Total closed shops.................... 19,668,793 28.5% Total membership (including open shops).... 23,839,906 23.5% Il-he reports from the Mciraw-Hill Publishing Company concern only its printing business. reFigures from the records of the Cost Department, New York Employing Printers' Association. 14 T4HE NEW YORK PRINTING INDUSTRY From this table it is seen that in comparison with the group of printers with which this investigation is most directly concerned, the 16 firms represent approximately one quarter of the mechanical payroll. The mechanical payroll is the commonest measure of the size of different concerns in trade parlance, and forms the basis for membership fees in the New York Employing Printers' Association. The two comparisons just preceding show that the data secured for this investigation represent approximately onetenth of the whole book and job branch of the industry in New York City, and about one-fourth of that portion of the industry affiliated with the New York Employing Printers' Association. With a quantitative showing no better than this, one is justified in asking whether the fact that these few firms are all much larger than the average printing concern would not tend to destroy their usefulness as typical indications of the industry's economic situation. It may be pointed out, for example, that all of these concerns have well planned cost systems, and that they are exceptional in this respect.'4 It may be argued that fluctuations in these firms would show proportionately less magnitude than in the many smaller concerns of the industry for the very reason that they are better able to plan their production. On the other hand it may be argued with equal force that fluctuations in large concerns must inevitably be larger than in small ones, because a skeleton complement of foremen and men constitutes a much larger proportion of the whole force in the small plant than in the large one, and the opportunity to vary costs to meet the pressure of workdemands is therefore greater in the larger firms. It is conceivable that these two situations cancel each other in effect.15 "This chart shows that there is [sic] in the City of New York. out of 797 commercial printing establishments, 56 plants operating their business off of a known cost of production. Gentlemen, that is a deplorable condition, 56 plants out of 797." Address of Hooper P. Hogan, reported in "Survey of the Printing Industry of New York City," the United Typothetae of America, January, 1920. This condition is well recognized in the industry. "See also w. I. King, "Employment, Hours and Earnings in Prosperity and Depression," 1923, pp. 31, 35. THE IMEASURES 15 But it is unnecessary to resort to speculation in answering this question. Fortunately two series of data are available which overlap the more complete series. In collecting data for the wage arbitrations of 1922, the Division of Industrial Relations of the New York Employing Printers' Association gathered payroll and billings figures from 40 firms in addition to those reported in this investigation. These 40 concerns are considerably smaller on the average than the 15 firms with which they can be compared.'6 As will be noticed in the tables, the total billings of the 40 firms constitute only about 60 per cent as much as the same total for the 15 firms in individual months. Tables 4 and 5 below have been drawn for comparison of the two groups of concerns. To make them directly eomparable they have been cast in terms of relatives on a base (100) which is the average of 1921-22. From these tables it is clear that the comparison of the billings of larger and smaller concerns is less conclusive than the comparison of their mechanical payroll fluctuations. A difference of more than 15 points occurs five times as between the two billings series, and not at all as between the two payroll series; a difference of more than ten points occurs 9 times in billings (32.2 percent of the cases) and only 4 times in payroll (14.3 percent of the cases); a difference of 5 points is exceeded 14 times (50 percent of the cases) in billings, and only 8 times (28.6 percent of the cases) in payroll. The average difference between the two billings series is 8.9, and only 4.7 between the two payroll series. '"In the list of the 40 firms which follows, those marked with an asterisk are the only ones approaching in the size of their business the 15 firms belonging to the other group: Acme Press- Afferton Press; Appeal Printing Co.; Art Color Printing Co. (*); Edward S. Benedict; brooklyn Eagle job Printing Department (*); Burr Printing House (*); S. J. Clark's Sons; Correct Printing Co.; Wm. arling Press; Herman Diamond, Inc.; Emporum Press; Herman Feinberg; Guide Printing & Publishing Co. (*) Kalkhoff Co., Inc. (*) Montague Lee & Co. Lent & Graff Co.; Libien Press; Lisso & Hartig; Jacob Luick; Marchbanks PressP ressPress; Ogilvie Press; Patteson Press; Peck & Durham; Price Bros. & Co.; John. Rankin Co. (*) Read Printing Co. (*) Redler, Inc.; Richardson Press; Rogers & Co.. Inc. (*); Wiiam E. Rudge,. Inc. (*)- Scribner-Press (*); Hiram Sherwood (*); William Segrist, Inc.; Stettiner Bros., Inc.; System Printing Co.; Ullman Press, Inc.; Charles P. Young & Co.; and Zeese-Wilkinson Co.; Inc. (*). 16 16 ~THE NEW YORK PRINTING INDUSTRY TABLE 4 17 COMPARISON OF THE BILLINGS OF 15 FIRMS WITH THE BILLINGS OF 40 FIRMS, SEPTEMBER, 1920, TO DECEMBER, 1922 15 firms 40 firms 15 firms 40 firms Date (dollars) (dollars) relatives18 relatives'8 Differences",(Average 1921-22=100), + I__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ I _ _ _ _ _ __ _ _ _ _ _ 1. _ _ _ _ _ _- _ _ _ _ _ _ _ 1920 Sept. Oct. Nov. Dec. 1921 Jan. Feb. March April May June uly Aug. Sept. Oct. Nov. Dec. 1922 March April May Aug. Sept. Oct. Nov. Dec. 51,305,170 1,133,415 1,032,871 1,240,828 1,224,674 944,216 992,221.992,224 838,673 902,241 815,810 909,152 835,106 854,724 891,185 1,193,447 843,372 872,557 885,731 929,422 976,239 856,164 733,787 818,784 819,188 899,954 899,690 1,059,285 $816,601 738,949 725,945 762,464 668,170 608,408 646,171 594,734 522,445 517,488 480,300 578,181 501,959 508,102 523,786 473,956 625,366 579,346 681,234 613,727 601,647 562,481 518,566 583,672 533,736 593,479 658,253 779,781 142 124 113 135 134 103 109 109 92 99 89 99 91 93 97 130 92 95 97 101 107 94 80 89 89 98 98 116 141 127 125 131 115 105 111 102 90 89 83 100 86 87 90 82 108 100 117 106 103 97 89 100. 92 102 113 134 1 48 12 16 5 20 5 9 11 3 4 15 18 "See also Chart 1. '"Not adjusted for seasonal variation. "'The average difference of the group of 15 firms from the group of 40 firms, in terms of relatives (disregarding signs), is 8.9; the median of the differences is 5.6. Ia 11 case the relative for the 15 firms is greater, in 17 cases smaller, than the relative for 40 firms. THE MEASURES1 17 TABLE 530I COMPARISON OF THE MECHANICAL PAYROLLS OF 16 FIRMS WITH THE MECHANICAL PAYROLLS OF 40 FIRMS, SEPTEMBER; 1920, TO DECEMBER, 1922. 16 firms 40 firms 16 firms 40 firms Date (dollars) (dollars) relatives21 relatives'1 Dif6fer-encesn ___________________(Average 1921-22 100) + - 1990 Sept. Oct. Nov. Dec. 1991 March ~Ayr June Aug. Sept. Oct. Nov. Dec. 19ff March April Sept. Oct. Nov. Dec. $515,333 508,187 499,640 526,223 542,903 451,361 479,421 445,236 437,734 440,044 422,312 487,137 436,312 469,698 481,791 514,196 487,014 433,141 524,550 458,693 502,607 481,372 446,133 453,056 459,121 499,011 494,170 524,113 $222,713 224,84 199,718 225,954 213,537 188,910 224,748 206,144 183,260 190,849 184,942 206,426 183,339 180,200 177,017 222,017 196,748 202,705 216,317 201,141 193,124 204,129 182,169 192,522 201,830 204,621 199,703 232,271 109 107 105 111 114 95 101 94 92 93 89 103 92 99 102 108 103 91 111 97 106 102 94 96 105 104 ill 112 113 100 113 107 95 113 106 92 96 93 106 92 90 89 111 99 102 108 101 97 102 91 96 101 103 100 116 7 0 0 0 4 3 6 12.. 12 0 3 4 3 0 4 "See Chart 2. "Not adjusted for seasonal variation. "The average difference of the group of 18 firms from the group of 40 firms, in terms of relatives (disregarding signs), is 4.7; the median of the differences is 4.0. In 10 case the relative for the 18 firms is greater than, in 13 cases smaller than, in 5.,cases the same as, the relative for 40 firms. 18 THE NEW YORK PRINTING INDUSTRY When the nature of the billings and payroll series is considered the reasons for the greater divergence of the former, as between large and small firms, are made apparent. Billings are made on the completion of individual orders, and the time element varies with the size of the orders. There is nothing to prevent billings from varying almost independently of time; the delivery of individual orders may or may not coincide with calendar intervals. Moreover, the more liberal use of cost accounting systems by the group of large concerns than by the smaller ones might be expected to give rise to CHART 1 THE BILLINGS OF 15 COMPANIES COMPARED WITH THE BILLINGS OF 40 COMPANIES, SEPTEMBER 1920 TO DECEMBER 1922, NOT ADJUSTED FOR SEASONAL VARIATION. Average 1921-1922 = 100. Unit, one percent. /-920 92/ /2Z THE MEASURES 19 divergences in the prices charged for similar work, and the price element bulks large in billings. But in the payroll series neither the time nor the price elements differ as between the big and small concerns. By agreement with the several unions involved23 and by trade practise, all companies pay wages weekly. And although a different proportion of the total number of employees are paid a premium over the currently agreed wage scale in different plants, the difference does not depend on the size of the plant. Thus the payroll series might be expected to show a fairly close relationship between large and small companies. CHART 2 THE MECHANICAL PAYROLL OF 16 COMPANIES COMPARED WITH THE MECHANICAL PAYROLL OF 40 COMPANIES, SEPTEMBER 1920 TO DECEMBER 1922, NOT ADJUSTED FOR SEASONAL VARIATION. Average 1921-1922 = 100. Unit, one percent. 12C~~~~~~' //O /00 A 41R 292 0 6~~~~~Y A more significan*t comparison between the series may' be made by examining the kind of divergence at the trough and ~~- -p - * — IO --- —- - - --- - -- -- /920 /92/ar A more significant comparison between the series may be made by examining the kind of divergence at the trough and during the revival. In the billings series the big companies did not decline as far, typically, as the small ones; and they lagged behind the small ones in revival. To be sure, the lowest point in either billings series was reached by the large concerns as 2"Payment of wages shall be made weekly in cash, not later than seventy-two hours after the close of the fiscal week, and not more than seven working days shall elapse between pay days." Shop Rules and Scales, December 1, 1921, page 18, paragraph 97. - 20 THE NEW YORK PRINTING INDUSTRY an isolated dip in July 1922. This presents the possibility that smaller firms made their serious adjustments in the summer of 1921, and may have profited from that fact in 1922, when the larger firms may have been forced to make even more drastic concessions.24 But the large companies laid off their spare employees more quickly than the small plants, and re-employed them a little sooner.26 In general, the correspondence between the two series is close enough to indicate that it will be safe to assume the representative character of the longer series which have been secured for the larger concerns if the limitations here noticed -are kept in mind.28 The payroll (and its corresponding physical measure, employment) may be expected to yield closer estimates of comparability than the estimates of receipts (and their physical parallel, production.) RELIABILITY In so controversial and contentious a field as the determination of wages, data concerning the ability of the industry to pay which have been derived from sources controlled by only one side must be examined in the light of their reliability. It is not inconceivable that distorted or manipulated reports might be supplied. In the present instance two reasons may be advanced for the belief that the individual reporting firms have not manipulated the figures prior to their collection. The first is that they had no understanding as to the manner in which the data were to be treated in the office of their association, and as a result would not know what sort of distortion they might practise even if they were so inclined. The second concerns the limit to which one sort of distortionmight be used without "See note by M. C. Rorty, p. 97, "Business Cycles and Unemployment," 1923, and W. I. King, op. cit., pp. 31-35. But King is not here concerned with sales. See also below, pp. 36, 40, 54, 69. See above, p. 14. #No objective measure of the limit of tolerance is at hand. The correlation coefficient for the two billings series is +.68, and for the two payroll series is +.70, but these can be interpreted only arbitrarily. THE MEASURES 21 detection. It would be to the advantage of the association in any single wage negotiation if it could show a decline of some kind, either in volume of manufacture (impressions), or in its billings. Over a longer period a manipulation of this sort would become perceptible as soon as the figures were out of line with employment information in the possession of the unions, or with the common knowledge of individual employees gleaned as a result of the number of hours which they might have worked in overtime. In addition, the figures which have been reported do not show evidence of manipulation.27 '271It must be obvious that any use of similar figures in the future, in wage negotiations, would need to be hedged about with suitable safeguards. The New York Employing Printers' Association is a fundamentally honest group; but in special instances trade unions, if they had no participation whatever in preparing the conclusions from such figures, would be justified in raising the question of bias. CHAPTER III STATISTICAL TREATMENT OF THE DATA This chapter will describe the manner in which it was necessary to treat the several series of fluctuations in order to render them comparable. In three cases the reporting concerns did not submit information recorded throughout the period by months. These conceirns reported their press impressions for the first quarter of 1920 as one figure, and this figure had to be distributed over the three months in as reasonable a manner possible. The basis of this distribution was the number of employees on the payroll of these concerns, taken separately. First, the ratio of the number of employees to the press impressions in later months was computed, and an average reached for these ratios. Second, the distribution of the quarterly figure for press impressions was made in the same proportion that held in the distribution of the number of employees monthly in the corresponding quarter. Finally, this distribution was examined to see whether the ratio of the computed monthly press impressions to the given monthly number of employees corresponded to the average ratio arrived at in the first transaction. In each case the correspondence was close enough to warrant calling the method reasonably accurate. This was the procedure in the three cases in which the data were reported as a quarterly total rather than as monthly.totals. This procedure obviously loads any comparison which may be made between press. impressions and the number of employees. But the loading has little influence, since the number of concerns reporting insufficiently is only 3 out of 17, and only one quarter is involved. 22 STATISTICAL TREATMENT 23 WEIGHTING THE PRESS IMPRESSIONS The impressions, as previously stated, were reported in three classes: those from platen presses, cylinder presses, and rotary presses. Each of these types of presses has a different labor complement and a different speed of operation. To add together the crude figures for the different types of press impressions would result in a haphazard weighting much like that involved in adding numbers of freight trains together without noting the number of cars carried by each. To render the press impressions comparable with a deliberate weighting it is necessary to consider the differences in labor complement and running-speeds. From the sources previously described' it was possible to secure equipment and cost information as follows: the number of presses of each type, their cost per hour, and the average number of impressions per hour for each type of press. Weighted averages of all of the platen, cylinder and rotary presses were computed as in Table 6, and the final weights used in summing the press impressions of different types were found by noting the relation of the platen average to the other two. The platen was weighted 1, the cylinder 2.3, and the rotary 2.11. So weighted, the different press impressions were added together and the monthly totals arrived at which are reported in Table 7. CHOICE OF A BASE The first problem in rendering the series comparable is the choice of a suitable base. Such a base must fairly represent the data, in that it must contribute to a clear interpretation of them. Several bases were examined from this standpoint, and were rejected. It was impossible to discover a secular trend from information extending over only four years, and arbitrary 'See page 11. TABLE 6 COMPUTATION OF WEIGHTS FOR DIFFERENT TYPES OF PRESS IMPRESSIONS 0_ _ TYPE OF PRESSES E *, 8 ~~~ o ~ ~ ~ M Platen, hand, lOx5 and smaller. 1,707 853 $1.24 $1.45 1,456,071 $2,111,302.95 Platen, hand,largerthan l0x15. 1,311 876 1.54 1.76 1,148,346 2,021,088.96 Platen, mechanically fed...... 625 1,388 1.61 1.16 867,500 1,006,300.00 Automatic................... 353 2,112 3.09 1.46 745,536 1,088,482.56 4,217,453 6,227,174.47 $1.48 1.00 Pony cylinders, hand.............. 333 1,192 2.94 2.46 396,936 976,462.56 Medium cylinders, hand....... 625 985 3.21 3.26 615,625 2,006,937.50 Large cylinders, hand......... 434 1,112 3.66 3.29 482,608 1,587,780.32 Cylinders, mechanically fed.... 282 1,022 4.74 4.63 288,204 1,334,384.52 Two-color presses............. 100 1,076 5.50 5.11 107,600 549,836.00 1,890,973 6,455,400.90 3.41 2.30 Webb (rotary).........110 2,873 9.00 3.13 316,030 989,173.90 3.13 2.11 _ _.. 0 3 z z M. 0 cn PI 13 Up4 STATISTICAL TREATMENT 25 allowances, derived from the growth of the industry as exhibited in Census figures, would have served only to confuse. The moving annual average was discarded because it involved losing six months at each end of the period, or one-fourth of the data. - Two arbitrary plans were hit upon, both of which answered the purpose. The first was a horizontal line at the arithmetic average of all the data; the second was a horizontal line at the arithmetic average of the figures for 1920-21. Both were used, but the first primarily. The second was used only in comparing printing production with the magazine advertising lineage figures of the Harvard Economic Service.2 The base of the latter series (the average of 1919-21) suffered less than 1 percent of change when altered to the average of 1920-21. The experiments which led to this choice were made only with the production series, and the base was applied to the other series without further experimentation. SEASONAL VARIATION The other- aspect of rendering these series comparable is the discovery and elimination of seasonal variation. The discovery of a seasonal swing might be expected to throw some light on the character of the industry, as well as on differences between the several series with which this study is concerned. The method used in determining the magnitude of the monthly seasonal indexes was substantially that described and used by the Harvard Economic Service," with one important modification. Because these series extended over only four years it was impossible to use the median link relatives of the successive items, as prescribed by the Harvard method. The median of only four items is an absurdity. The difficulty was here evaded by using, as the typical representative of the 'See below, p. 58. "'Correlation of Time Series," by W. M. Persons, Quarterly Publications of the American Statistical Association, June, 1923, pp. 714-721. 26 THE NEW YORK PRINTING INDUSTRY link relatives, a weighted average in which the two central link relatives were weighted 4 each and the two outside link relatives 1 each.4 For instance, the following table shows the computation of the typical link relative for the January items: PRESS IMPRESSIONS January Link relatives Weights Products 1921 107.4 1 107.4 1922 92.5 1 92.5 1923 99.2 4 396.8 1924 98.5 4 394.0 10 990.7 Typical Link Relative = 99.1 The typical link relatives for the other months were computed in the same fashion and the prescribed adjustments and transactions were made on the basis of them. In eliminating the seasonal variation from the several series, the base line ordinate (in these series, since the slope is zero, the ordinates for the base line are the same throughout) was multiplied by the successive seasonal indexes. The original items were then divided by their corresponding figures on this adjusted base, and the resulting quotients formed the relatives, adjusted for seasonal variation, on the 1920-23 base. No computation of the seasonal variation for purchases was made, for purchases are made independently of seasons. The profit and loss items showed a distinct seasonal swing, but since both positive and negative items (losses) entered into the series a roundabout plan had to be adopted in expressing them as relatives. First, the total cost series and the billings series were both expressed as relatives on the base of the January 1920 total cost figure ($975,306) as 100 percent. The total cost relatives were then subtracted from the billings 'This device was suggested by Dr. Frederick R. Macaulay. The weighting is arbitrary. 'The adjustment of the chained typical link relatives was made arithmetically not geometrically. See W. M. Persons, op. cit., p. 717, note. STATISTICAL TREATMENT 27 relatives, and the difference was added (algebraically) to 100. For instance, the total cost figure for February 1920 ($861,395) and the billings figure for the same month, for 13 firms,($1,161,713) were expressed as percentages of the January 1920 total cost figure. These relatives were 88.32 and 119.11 respectively. The difference between these relatives is 30.79, and when 100 is added the figure becomes 130.79. This is the crude profit relative for February 1920. The other items were computed in the same way. This resulted in a series which was treated as original items, for, when plotted, its curve was identical with that of the original profits series except for the scale of ordinates. It was adjusted for seasonal variation and expressed as adjusted relatives on the average of 1920-23 as 100 per cent, like the other series. This scheme was the invention of Mr. C. A. Hokinson. In a sense it misrepresents the -situation, for the change of scale resulting from the roundabout calculation seems to give it a narrower range of fluctuation, whereas in fact profits fluctuated through a greater range than any of the other series. But if the scale distortion is kept in mind it should not be misleading. CHAPTER IV FLUCTUATIONS OF PRODUCTION, EMPLOYMENT, SALES, AND PAYROLL In Table 7 are recorded the original figures on which the investigation is based. (Press impressions are recorded after weighting.) By the methods described, these figures have been reduced to relatives on the 1920-23 base, adjusted for seasonal variation' but not for secular trend, and the resulting relatives are reported in Table 8. Table 9 contains the various indexes of seasonal variation. The task of this chapter is to see how much light can be thrown on the economic conditions of the industry by the use of these figures. The first problem is the examination of the data from the point of view of their consistency with each other and with what is commonly known about the industry. Are the different series consistent? Are the inconsistencies greater than one might expect to arise from faults of method, and do they indicate significant aspects of the industry beyond their several margins of error? What features of the industry are thrown into discernible relief? In answering these questions, it will be convenient to study the indexes of seasonal variation, and then the first five2 of the fluctuating measures, to see how they differ in timing and intensity. SEASONAL VARIATION Although the magnitude of the seasonal swing might be expected to differ as between the several series, its direction EExcepting purchases. 'Analysis of the overhead, total costs, and profit and loss items will be reserved for eparate chapters. See Chapters V and VI. 28 PRODUCTION, EMPLOYMENT, SALES, AND PAYROLL 29 TABLE 7 THE DATA Press impres- No. of Pur- Over- Total Profit sions em- Billings Payroll chases head costs and loss Date 17 firms ployees (dollars) (dollars) (dollars) (dollars) (dollars) (dollars) (000 15 15 firms 16 firms 14 firms 13 firms 13 firms 13 firms '' omitted) frms 1920 Jan. Feb. Mar. June uly Aug Sept. Oct. Nov. Dec. 1921 'Lan. Feb. Mar. 2up, J une uly Aug. Sept. Oct. Nov. Dec. 1922 an. Mar. Aug. Sept. Oct. Nov. Dec. 1923 Jan. Feb. Mar. June July Aug. Sept. Oct. Nov. Dec. 200.403 176,485 209.504 164,025 166,116 165,737 170,331 182.484 168,75 159,459 140,187 147,728 158,660 110,358 120,636 114,495 108,246 103,813 109,092 122,834 103,853 106.405 121,890 127,231 117,682 107.173 124.381 124.361 121,608 120,363 121,044 123,307 118,062 132,242 128,421 139,413 138,202 117,254 135,633 126,817 130,959 126,319 121,044 126,290 116.059 152.779 138,506 126,368 I 3,293 3,295 3,220.2,918 2,768 2,780 2,770 2,973 2,780 2,59 2,8 2,678 2,601 2,445 2,193 2,163 2,117 2,211 2,214 2,370 2,348 2,301 2,393 2,500 2,452 2,573 2,413 2,397 2,372 2,299 2,403 2,480 2,644 2,404 2,460 2,677 2,522 2,553 2,394 2,452 2,537 2,381 2,381 2,241 2.361 2,583 2,525 2,576 I $1,197,066 1,211.659 1,534,990 1,276.282 1,215,061 1,214,086 1,166,193 1,195,428 1,305,170 1,133,415 1,032,871 1,240,828 1,224.674 944,216 992,221 992,224 838,673 902,241 815,810 909,152 835,106 854,724 891,185 1,193,447 843,372 872,557 885.731 929,422 976.239 856,164 733,787 818,784 819,188 899,954 899,690 1,059,285 969,993 846,705 1,000,941 966,319 1,087,681 947,251 879,375 837,476 869,585 997.790 1,063.354 1,066,101 I I $584,597 547.670 649,846 562,575 526,292 550,524 510,957 558.937 515,33 508,187 499,640 526,223 542,903 451,361 479,421 445,236 437,734 440,044 422,312 487,137 436,312 469,698 481,791 514,196 487.014 433,141 524,550 458,693 502,607 481.372 446,133 453,056 459,121 499,011 494,170 524,113 569,719 473,345 542,299 506,278 542,814 484,680 472,462 503,785 476.229 579.363 550,206 541,821 I $332,842 257.57C 324,644 252,18Z 371,027 407,031 375,930 387.232 402.752 378.822 223,522 231,903 199,467 154,646 171,329 171,248 145,428 190,226 168,710 170,049 194,329 227,048 208,562 181,822 193,50 169,933 227,2 216,017 221,347 231,888 167.981 195,317 196,762 245,132 242,281 265,139 243,896 196,405 256,781 196.290 270,993 190,122 187.693 198,713 187,565 247,193 230,903 262,057 I I I $195,443 181.392 198,426 201,491 205,469 211,003 205,591 197,649 219,383 229,876 220,948 290,425 225,145 212,091 222,473 216,921 205,879 216,922 212,391 207,914 222,909 220,489 218,609 266,642 235,104 222,500 236,012 214.281 225,743 224,121 208,970 212,753 232,873 233,564 232,670 258.201 234,875 222.820 239,909 236,194 240.450 245,040 242,357 229,528 222,159 234.297 244,428 229,595 I I $975,30( 861,391 1,051,171 906,26E 981,520 1,060.971 985,681 986,921 1,023,124 1,013,92E 828,54C 935,282 827,960 706,061 773,785 739,053 686,924 758,655 717,413 712,865 764,162 814,723 823,433 859,437 807,738 723.281 868,287 791,064 837,247 847,181 714,177 752,733 774,674 860,908 867,833 923,987 915.356 779,517 924,178 818,359 946,391 820,776 789,348 832,263 782,765 934,434 692,679 914,203 i $100,115 ) 300.318 i 380.434 323,237 207,611 125,665 I 147,904 167,582 148,009 67,455 127,719 189,590 261,866 92,420 143,016 177,134 101.952 93,808 63,465 41,057 52.393 1,381 27,966 275,930 1,264 (loss) 65,475 34,867 (loss) 92,235 88,318 45,065 (loss) 39,082 (loss) 16,494 (loss) 34,436 (loss) 71,778 (loss) 35,563 (loss) 36,760 51,641 (loss) 11,305 261 96.291 50,308 80,195 50,614 35,490 (loss) 42,661 501 (loss) 282,159 58,414 I I I I I I I -- I ' ' 30 THE NEW YORK PRINTING INDUSTRY in each case (except overhead) ought presumably to be the same. With the exception of the employment series and the overhead series there are few differences in direction. These TABLE 8 ADJUSTED RELATIVES. AVERAGE 1920-23=100. Press No. of Purchases Date impres- em- Billings Payroll 14 firms Overhead Total Profits' sions ployees 15 firms 16 firms (unad- 13 firms costs 13 firms 17 firms 15 firms justed) 13 firms 1920-Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. 1921-Jan. Feb. Mar. June July Aug. Sept. Oct. Nov. Dec. 1922-Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. 1923-Jan. Feb. Mar. Tune July Aug. Sept. Oct. Nov. Dec. 145.0 143.1 147.2 124.2 126.4 129.0 126.6 127.4 127.7 112.2 103.5 104.9 114.8 89.5 84.8 86.7 82.2 80.8 81.1 85.8 79.6 74.8 89.8 90.3 85.1 86.8 87.5 94.1 92.4 93.8 90.0 86.0 90.4 92.9 94.6 99.0 100.0 95.2 95.2 96.0 99.6 98.3 90.0 88.2 88.8 107.3 102.0 89.7 125.0 127.0 131.2 119.6 116.3 117.3 112.5 116.3 105.9 99.0 96.1 97.4 98.7 94.2 89.3 88.6 89.0 93.3 89.9 92.7 89.5 90.8 94.0 91.0 93.1 99.2 98.3 98.2 99.7 97.0 97.6 97.0 100.7 94.9 96.6 97.4 95.7 98.4 97.5 100.5 106.6 100.5 96.7 87.7 90.0 102.0 99.2 93.7 117.0 127.2 144.0 123.4 118.7 124.8 130.9 125.5 135.6 111.8 100.9 107.3 119.7 99.1 93.1 95.9 81.9 92.7 91.6 95.5 86.8 84.3 87.1 103.2 82.4 91.6 83.1 89.9 95.4 88.0 82.4 86.0 85.1 88.7 87.9 91.6 94.8 88.9 93.9 93.4 106.3 97.4 98.7 87.9 90.4 98.4 103.9 92.2 106.5 115.0 121.3 115.9 106.7 112.3 109.7 110.5 108.6 98.2 98.1 98.2 98.9 94.8 89.5 91.7 89.3 94.4 90.6 96.3 91.9 90.8 94.6 96.0 88.7 91.0 97.9 94.5 101.9 98.2 95.7 89.6 96.7 96.5 97.1 97.8 103.8 99.4 101.2 104.3 110.0 98.9 101.4 99.6 100.3 112.0 108.1 101.2 140.5 108.8 137.0 106.5 156.6 171.8 158.7 163.5 170.1 159.8 94.4 97.9 84.2 65.3 72.3 72.3 61.4 80.3 71.2 71.8 82.0 95.8 88.1 76.8 81.7 71.7 95.9 91.2 93.4 97.9 70.9 82.5 83.1 103.5 102.3 111.9 103.0 82.9 108.4 82.0 114.4 80.3 79.2 83.9 79.2 104.4 97.5 110.7 87.1 86.3 88.6 93.4 94.7 95.3 94.9 95.7 99.1 102.1 99.3 111.0 100.3 100.9 99.3 100.7 94.9 97.9 98.1 100.6 100.7 98.0 98.2 101.9 104.8 105.9 105.4 99.4 104.0 101.2 96.5 103.0 105.2 103.8 104.5 98.7 104.7 106.0 107.1 109.6 110. 110.6 111.9 111.1 100.4 104.1 109.8 87.7 112.5 114.5 118;4 114.6 116.8 122.6 121.7 119.5 120.5 109.6 100.1 101.5 95.5 93.8 87.1 93.5 81.7 87.7 88.6 86.3 80.0 88.0 99.5 93.3 93.2 96.1 97.8 100.1 99.6 97.9 88.2 91.1 81.2 93.0 104.8 100.3 105.6 103.6 104.1 103.5 112.6 94.9 97.4 100.7 92.2 101.0 83.7 99.2 105.3 113.9 120.3 111.3 103.7 100.7 104.2 109.0 107.3 105.4 103.7 103.8 121.1 95.5 99.2 94.2 94.0 97.7 96.3 97.0 98.4 98.6 94.3 112.0 95.5 93.1 83.4 91.6 93.3 85.0 86.8 91.5 90.0 91,4 88.3 90.2 90.5 96.9 95.2 100.3 98.0 105.3 104.2 99.1 106.7 108.3 127. 2 100.7 'See above, pp. 26, 27. PRODUCTION, EMPLOYMENT, SALES, AND PAYROLL 31 TABLE 9 INDEXES OF SEASONAL VARIATION Produc- Employ- Over- Total tion ment Billings Payroll head Costs Profits Jan. 102.2 104.2 101.8 109.1 101.0 102.2 94.1 Feb. 91.2 102.6 94.7 94.6 94.6 88.7 103.2 March 105. 2 97. 1 105. 9 106. 5 100. 8 104. 7 103. 8 April 97.7 96.5 102.8 96.5 97. 0 93.2 107.4 May 97.2 94.1 101.8 98.0 97.7 99.1 105.6 June 94.9 93.7 96.7 97.4 99.7 102.0 100.9 July 99.4 97.4 88.6 92.6 97.5 95.5 99.3 Aug. 105.9 101.1 94.7 100.5 93.0 97.4 96.6 Sept. 96.6 103.8 95.7 94.2 99.6 100.1 96.4 Oct. 105.2 100.2 100.8 102.7 101.3. 109. 1 91. 1 Nov. 100.3 100.7 101.8 101.1 100.0 97.6 98.1 Dec. 104.2 108.7 115.0 106.4 117.8 108.6 103.5 two show a number of idiosyncrasies. Employment diverges from the others in March, July, and October; and from the payroll series, with which it might be expected to run most closely, it diverges also in May and September. The explanation may lie partly in overtime and short time, but it is probably also true that the method used in segregating the seasonal fluctuations is not sufficiently exact, since only four years of information were available. The latter also is probably a partial explanation of the independence of the overhead swings, although they cannot be counted on to run consistently with the others because of the peculiar nature of overhead costs. Indeed, it is precisely because of their independence that they are called "overhead costs." Because of the difficulty in computing exactly the seasonal indexes, it will be necessary to refer occasionally to the unadjusted data (Table 7) in arriving at a plausible explanation of otherwise puzzling situations. But the seasonal indexes and the relatives derived from them will be used, in default of anything better, for comparing these series with others not directly connected with the industry. 32 THE NEW YORK PRINTING INDUSTRY In general, so far as can be judged from the four years available, the typical printing year is one which starts well in January, falls off slightly in February, recovers in March to a slight extent only to fall off once more during the summer months, and finally revives in the last three months to its annual peak in December. Production, employment, sales (billings), and the four cost series may be compared with reference to their timing and intensity. Any wide inconsistencies which appear from these comparisons will need to be explained; and out of these explanations may arise some hints regarding peculiarities of the industry. By what criterion shall one judge whether the consistency is sufficient to warrant drawing conclusions about the economic condition of the industry? A difference of two months in timing will need close examination: if one of the series differs by more than two months in rise or fall, in peak or trough, from any other (purchases and overhead excepted) that difference will be more than one would expect. The case of overhead will be examined more carefully in another chapter. Purchases are affected by an opportunistic influence from which the other series are exempt, for purchases are made in part because they are needed, and in part because market conditions are favorable. The chance for forward buying at a time of rising commodity prices, and the chance for storage against immediate need, take the purchases out of immediate contact with production, although one might count on a vague general correspondence between the two. In intensity, for the same reasons, purchases will probably vary much more widely than the rest, simply because they can and because the others are relatively fixed. Billings will fluctuate less than some others, because they depend in part on contracts between printer and customer, contracts in which prices are fixed for units of production. No such element is involved in the production and ASee Chanter V. PRODUCTION, EMPLOYMENT, SALES, AND PAYROLL 33 employment figures, which may accordingly be expected to fluctuate more widely. With so large a use of machinery, one would expect employment to fluctuate less widely than production. The payroll series ought to vary least of all, partly because wage rates are fixed for long periods by contract with the unions of the industry, partly because of the common practise of retaining the more skilled and consequently better paid workers when a lay-off becomes necessary. TIMING Three points may be noted in the timing of the several series: the date of the peak, the date of the trough, and the elapsed interval between peak and trough. Table 10 contains these comparisons. The erratic behavior of the purchase figures has been discussed, and their influence on the total costs, in which they are included, is so great as to prevent the total costs from having much relevance to the present question. The relevance of total costs to the question of profits will appear later,5 and they can be usefully omitted from this discussion. The dates of the overhead peak and trough show that the very direction of the overhead swing was opposite to that of the others, and makes a separate discussion desirable. The fact that the peak dates (relatives) of all four of the other series are alike says a good deal for the validity of the seasonal indexes, but the fact that the trough dates (relatives) are all different says more to deny that validity. The series reached the trough dates (relatives) in the following order: 1. Employment 2. Billings 3. Production 4. Payroll This order may be partly explained by the following considerations. In reducing costs in an industry in which the wage scale is fixed from year to year, the first point of attack 'See Chapter VI. Ct WC TABLE 10 COMPARISON OF THE TIMING OF THE SERIES.6 -Production Employm't Billings Payroll Purchases Overhead Total Costs Peak date (relatives). Mar., 1920 Mar., 1920 Mar., 1920 Mar., 1920 June, 1920 Dec., 1920 June, 1920 Peak date (actual).... Mar., 1920 Feb., 1920 Mar., 1920 Mar., 1920 June, 1920 Dec., 1920 June, 1920 Trough date (relatives) Oct., 1921 April, 1921 May, 1921 Jan., 1922 May, 1921 Feb., 1920 Sept., 1921 Trough date (actual).. June, 1921 May, 1921 July, 1922 July, 1921 May, 1921 Feb., 1920 May, 1921 Elapsed intervals (relatives)......... 19 mos. 13 mos. 14 mos. 21 mos. 11 mos........... 15 mos. Elapsed intervals (actual)........ 15 mos. 15 mos. 28 mos. 16 rnos. 11 mos............ 11 mos. The relatives and actual dates do not always correspond because the relatives have been adjusted for seasonal variation, which is ignored in the actual dates TABLE 11 COMPARISON OF THE INTENSITY OF FLUCTUATIONS IN THE SERIES Production Employm't Billings Payroll Purchases Overhcad Total Costs Range (relatives)..... 147.0 131.2 144. 0 121.3 171.8 111.0 122. 6 74.8 88.6 81.9 88.7 61.4 86.3 80.0 72.2 42.6 62.1 32.6 110.4 24.7 42.6 Percent drop (actual).. 59.2 32.8 52.2 24.9 64.4 37.8 35.2 TJ z 03 z z C4 0 IV pi "IO C0 PRODUCTION, EMPLOYMENT, SALES, AND PAYROLL 35 would naturally be on employment. Since the cheaper employees would be paid off first, the payroll would not be affected in the same degree. If price-cuts were extensive some business might have been retained, so that billings touched bottom ahead of production. The payroll withstood the depression longest and did not reach its trough until production had begun to revive. When the actual trough dates are considered, the order is as follows: as follows: 1. Employment 2. Production 3. Payroll 4. Billings Production and payroll stand in the same relation to each other and to employment. The only change is the billings shift. This is a significant shift. Billings (actual) experienced a serious slump for the three consecutive months of July, August, and September, 1922. The fact that the slump covered three months, and the probability that the large plants which are represented in these figures took their losses fairly late in the depression,7 indicate that the seasonal indexes and the relatives which they affect tend to conceal the actual situation, so far as billings are concerned. The order of the elapsed intervals (relatives) is the same as that in which they reached their respective troughs, since they started together. But the intervals differ by a longer interval than the tentative criterion allowed. How can this be explained? It does not stretch the imagination too far to argue that the maximum difference (8 months between employment and payroll) is within reason, when the nature of the two series and the facts of the industry behind them are considered. A relatively large proportion of the employees of the industry are paid a premium over the minimum scale requirements, 'See above, p. 20. 36 THE NEW YORK PRINTING INDUSTRY principally because they are more efficient than the rest; and these employees are commonly retained in a depression. Accordingly the payroll is not affected by a lay-off to the extent that the employment figures are affected. If the men of greater skill are retained, the decline in production is subject to a similar delay, and this effect is heightened by their desire to prove their skill by efficient work. As it appears to them that no more are being laid off, their efforts slowly relax; and it is even conceivable that their incentive towards individual efficiency would fall away still further with the revival in employment. Billings are the product of production and price, yet the actual billings figures were nearly twice as long in touching bottom as the production figures. Unfortunately, there are no price data available with which to prove the only plausible explanation which presents itself. That is the one which was given above, that price-cutting was considerably delayed in the big companies. In preparing the brief for the employers in the 1922 arbitration a number of letters were secured from various printers, who reported in the first two weeks of November of that year that their prices had declined varying amounts. The most frequently mentioned decline was 20 percent. The smallest mentioned was 10 percent, the largest about 30 percent. These data are unsystematic and were taken from both large and small companies, so that they are not adequate evidence for the conclusion. But it is interesting to note that one of the largest companies reported, even in November 1922, the smallest decline in price, and no large company was included among those reporting the heavier cuts. A larger proportion of big companies than of small ones work on a contract basis. INTENSITY These fluctuations may also be discussed from the point of view of their relative intensities, with the use of two measures. The first is the range of the fluctuations of the relatives from PRODUCTION, EMPLOYMENT, SALES, AND PAYROI, 37 high point to low, and the second is the percentage drop of the actual (unadjusted) figures from peak to trough. Table 11 states these comparisons. The purchases again fluctuated most widely, as anticipated. Overhead shows the least violence among the relatives and occupies a middle position among the actual percentage drops. Total costs, probably because they include the widest and the narrowest fluctuations (purchases, payroll, and overhead), occupy a position midway between the extremes. The others show the following order of intensity: 1. Production 2. Billings 3. Employment 4. Payroll This order refers to the relative figures, and is changed, as in the timing order, when the seasonal indexes are disregarded. Production, employment, and payroll retain their relative positions, but billings show a greater intensity than any of these. Having denied the validity of the seasonal index for the billings series while discussing its timing, one must also deny it here and conclude that the unadjusted figure more accurately expresses the situation. Billings, then, fell lower than production, as well as falling later. The explanation is the same, that is, when the price cuts were finally made they were deep ones. These figures may also indicate that the New York printers were successful in retaining only the cheaper grades of work, allowing the rest to leave the city to competitors in other localities. No admissible evidence on this point is available. If the best men were retained, why did production decline more than employment? The opposite would appear to hold. The answer must be that the influence of short-time employment, and the abandonment of overtime, which would retain the names on the employment lists without retarding the pro 38 THE NEW YORK PRINTING INDUSTRY duction decline to the same extent, is responsible. No figures are available relating to the amount of short-time employment used in the industry, and accordingly this explanation is only conjectural. If the validity of these hypotheses be granted, there is no reason to discredit the assumption that the data are internally consistent. Some conclusions of interest regarding the industry may be drawn from these comparisons of timing and intensity. 1. Labor suffered more from unemployment (and possibly underemployment) during the recession of 1920-21 than from reductions in wages. This appears as the result of comparing payroll and employment, and concerns only the labor of the industry considered as a whole. Consideration of separate crafts might easily show the opposite to be the case. In some cases the ratio of payroll to employment increased. Indeed, this is precisely the effect of the agreements with the trade unions involved, since they establish a fixed minimum scale which can be changed not oftener than once a year. Men are e thrown out of work rather than reduced in wages, or they are employed part-time. The effect is made clearer by considering the manner in which the wage burden of the industry increased during the whole period 1920-23. Table 12 shows the wage changes in the industry. In this table the membership of the unions is taken at a given point and does not vary throughout the computation. Hours worked are ignored. The variable is the scale in dollars per week, whether the standard printer's week of 48 hours to May 1921 or the week of 44 hours after that month. The table shows that at no time since January 1920 have the wage rates been lower than they were at that time, and the alternative to the employer seeking to reduce his labor costs is the lay-off. This runs contrary to many employers' notion of the most desirable way of handling labor matters in time of depression. But it bears out the understanding PPRODUCTION, EMPLOYMENT, SALES, AND PAYROLL 39 TABLE 12 THE WAGE BURDEN OF THE BOOK AND JOB INDUSTRY, CONSIDERING ALL UNIONS AND ALL RATES.8 Date Weighted total wage scales Index January, 1920........................ $711,050 100.00 October, 1920........................ 782,150 110.00 April, 1921........................... 750,500 105.55 July, 1921............................ 742,100 104.37 December, 1921...................... 729,250 102.56 January, 1923.......................... 750,650 105.56 February, 1923...................... 753,800 106.01 September, 1923...................... 765,200 107.62 December, 1923..................... 793,100 111.54 January, 1924......................... 796,850 112.07 of some economists that an attack on the rate is viewed, in organized labor circles, with more serious alarm than a reduction in employment.9 2. The efficiency of printing labor increases so long as security of tenure is threatened. In the face of an impending lay-off the typical worker applies himself more closely. Members of Typographical Union No. 6 are exceptions to this situation from 1924 on, for the contracts with that union contain a seniority provision which provides that compositors. must be laid off in the reverse order of employment rather than because of an efficiency rating. Yet even in this case the rule applies in part, for any notorious inefficiency can be made the basis of discharge. The evidence for the conclusion that efficiency increases while employment is declining is found in the fact that production declined more slowly than employment. Fewer men were required for the same amount of IThis table was compiled by Mr. C. A. Hokinson on the basis of the changes in the actual rates and on the total membership of unions at a given time. These memberships are: Typographical No. 6, 6,500; Cylinder Pressmen No. 51, 2,800; Cylinder Feeders No. 23, 2,$00; Job Pressmen No. 1, 600; Job Press Feeders No. 1, 600; Paper Cutters No. 119, 900 Paper Handlers No. 1, 400; Sheet Straighteners No. 119, 600; Mailers No. 6, 200; Bindery Women No. 43, 2,100. 9"If the alternative to a temporary reduction in the rate is a reduction in output and less work, they explicitly prefer that alternative." Henry Clay, "The War and the Wage-earner," uoteW by Douglas, Hitchcock and Atkins, "The Worker in Modern Economic Society," p. 209. 40 THE NEW YORK PRINTING INDUSTRY work. An alternative interpretation is that the number of hours worked was being reduced simultaneously with working forces, their efficiency remaining constant. Although no information concerning the time worked is available, this explanation is less plausible than the other, for the employer probably does not use part-time work until late depression, when his force has reached a durable minimum. What happens to labor efficiency when employment begins to revive is not clear from these data. Production continued to decline after employment touched bottom. This may mean that employers were now resorting to short-time employment, or it may mean that employees, noticing the cessation of lay-offs, began to feel more secure in tenure and allowed efficiency to wane. Since the employment revival began quite cautiously, the first explanation is the more plausible. Moreover, production declined farther than employment. This points again to short-time work, for the only alternative hypothesis must be that a decline in efficiency, of greater influence than the minor increases in employment, took place; and it is at least doubtful whether workers are so promptly sensitive to almost imperceptible changes of this sort. All this is pure conjecture, for the essential data concerning hours worked are not available. 3. The behavior of the billings series presents an interesting situation. It declined farther than production, and later. As previously noted, contract work forms a large part of the production of big printing concerns, and this fact tends to delay price-cutting. At the expiration of contracts of a year's duration, the current situation in the industry may have changed in such a way that the printer is at the mercy of his customer in arranging a new contract. CHAPTER V FLUCTUATIONS IN OVERHEAD COSTS Since an unusual meaning is here given to the word overhead, it will be convenient to review its definition. As used in this study, overhead costs are all the costs not directly attributable to individual items of production. The only attributable items are those called mechanical payroll and purchases. Overhead includes all the other costs of the "sold product," except interest on investment, which printers call profit. In comparing this use of the term with the more familiar classification of the so-called "9-H" sheet of the United Typothetae of America, the printer should understand that overhead costs as here used comprise the following: office payroll, sales payroll, rent, heat, light, power, insurance, taxes, depreciation, bad debts, spoiled work, department direct expense, office stationery and postage, advertising, cartage and carfare, miscellaneous expense, selling expense, stock-handling expense, shipping, and office expense. This is different from the accountants' handling of the same items in that it includes items which accountants classify under separate headings. The excuse for tampering with the usual classification of the professional accountant is that this study is not primarily concerned with the cost accountant's problem, but rather with an economic problem, and a more general summary simplifies the task considerably. In order to show the relation of the different items listed above to the total costs in the average firm, it is necessary first to determine how these firms compare in size with those reporting to the Cost Commission of the United Typothetae of 41 42 THE NEW YORK PRINTING INDUSTRY America. That Commission's report, "Composite Ratios of Printing Costs, Analysis of 1922 Composite Statement," may then be used to show what proportions of the total cost are represented by each item above. TABLE 13 COMPARISON OF THE SIZE OF THE FIRMS REPORTING TO THIS INVESTIGATION WITH CERTAIN SIZE-CLASSES OF THE U. T. A. COST COMMISSION, 1922.1 Mech. payroll Sales Employees average average average Class G........ $180,846 $563,397 124 Class H............... 627,091 1,853,562 443 15-17 firms 347,541 705,806 164 Table 13 shows that the firms represented in this study are more nearly the size of the Class G firms of the United Typothetae of America report than of the Class H companies. The cost ratios which apply to the Class G companies will apply more closely than those of the Class H companies in showing the probable ratios of the firms reporting to this investigation. The reason for noticing the proportion of the total cost of "sold product" which is represented by the individual overhead items is that overhead fluctuates partly with and partly against the trend of the fluctuations of all the other series, and it is important to analyze the overhead costs as far as possible to see what explanation can be given for the shifting direction. Overhead costs, as reported in this investigation, were given only in lump sums, without analysis. To analyze these figures is therefore impossible. The next best analysis is that of the figures which are most comparable. The most comparable figures (those with which the proportions may be imputed to the figures for firms represented in the present study) are shown 'Computed from data furnished by the Cost Department, N. Y. Association of Employing Printers, and from Table II-A, "9-H Averages of Classified Expense Items in Dollars and Cents," Composite Ratios of Printing Costs, Analysis of 1922 Composite Statement, prepared by Accounting Bureau, Department of Research, United Typothetae of America. OVERHEAD COSTS 43 by Table 13. Table 14 gives the ratios of different overhead costs which will serve in part to explain overhead fluctuations. But the U. T. A. figures must be discounted in part because they are not typical of New York alone, since they cover the entire U. T. A. membership, and they are not restricted to firms whose output is largely the printing of periodicals. With these qualifications, Table 14 shows the cost ratios which may be applied to the overhead costs reported to this study. TABLE 14 PROPORTION OF TOTAL COST OF SOLD PRODUCT REPRESENTED BY DIFFERENT CLASSES OF EXPENSE, CLASS G, U. T. A. COST ANALYSIS,2 AND PROPORTION OF OVERHEAD ITEMS REPRESENTED BY CERTAIN CLASSES. Percentage to Percentage to Actual overhead (in total cost of present meaning) sold product Total "fixed" expenses.. $53,127 29.5 9.5 Total "current" expenses less wages........... 23,567 13.4 4.2 Total "general" expenses 55,724 31.5 10.0 Total "selling" expenses. 45,348 25.6 8.2 Total overhead (present meaning).......... 177,766 100. 0 (31.9) Total payroll.......... 180,846.... 32.5 Total purchases........ 197,994.... 35.6 Total, sold product550 556,606.... 100.0 OVERHEAD IN DOLLARS With these considerations in mind, the fluctuations of the lump sum of overhead costs can be reviewed. They are found in Tables 7 and 8, and Chart 3. The striking feature of the series is that its low point is within one month of the beginning the series. Starting low, it climbs rapidly until the end of the first year, and continues climbing irregularly until the end of the whole period. The other series all dropped from an early peak until the middle 'Composite Ratios of Printing Costs, Analysis of 1922 Composite Statement. 44 THE NEW YORK PRINTING INDUSTRY or end of 1921, and then recovered gradually until December, 1923. While production was falling off, overhead was mounting. When production had begun to revive, the rate of overhead's upward sweep had lessened, and with some exceptions overhead moved with production, responding to minor fluctuations. In December a considerable increase is noted except for 1923, and even in 1923 an increase was recorded over October. The seasonal index for December was greater than for any other series. CHART 3 FLUCTUATIONS OF PRESS IMPRESSIONS, OVERHEAD COSTS, AND TOTAL COSTS, 1920-1923, ADJUSTED FOR SEASONAL VARIATION. Average 1920-1923 = 100. Unit, one percent. u /92J /22 19f These facts raise several questions. What elements of overhead cause the lump sum to respond to minor fluctuations in production, and what elements cause the lump sum to oppose the major fluctuation downward in 1920-21? Why should overhead move against the tide at one period, and accompany it at another? What parts of overhead are relatively fixed, OVERHEAD COSTS 45 and what elements are variable and adaptable to production requirements? Clearly the readily modifiable elements in a book and job printing business are payroll, purchases, and some of the overhead expenses. Among these are light, power, spoilage, and department direct expense (oil, waste, proving paper, etc.); some part of office salaries; salaries and commissions of the selling force; and probably also the small items of bad debts and shipping expense. The fixed elements are rent, insurance, taxes, and depreciation. The fixed expenses must be met whether any business is transacted or not. The adaptable expenses may be varied tosuit the needs of the business. But it is important to notice that not all of the adaptable expenses vary in the same direction. In time of depression, the wise manager will curtail his office force slightly, but he will spend more money in the salaries of salesmen and in advertising, in an effort to retain some of the printing which is slipping to competitors. It is this last item which is probably responsible for the increase of overhead during the year 1920. Selling expense constitutes a good quarter of the overhead lump sum. Together with the fixed expenses (which cannot be reduced), selling expense constitutes 55.1 percent of overhead. The part of overhead which can be reduced consists of current expenses and general expenses, 44.9 percent of the total sum. If this were plausible, it could be inferred that in January 1921 the managers of these companies saw no sense in throwing more good money after bad, and so the rate at which overhead was increasing began to decline. With a fixed sales force and a set of fixed expenses, both amounting to 55.1 percent of the lumped overhead, the other adaptable costs might be expected to vary roughly in accord with production; which they did. How can the strong seasonal index for December be ac — counted for? Part of the -seasonal swing is due to the 46 THE NEW YORK PRINTING INDUSTRY increase of those items which tend to vary directly with production; but this does not explain the whole swing, which is greater than that of production. Two further explanations were advanced by Mr. C. A. Hokinson. The first was that in the last month of the calendar year business men realize that tax returns must presently be made, and make advance purchases when possible, accepting bills to reduce the year's profit showing. The second was that firms which are faced with serious tax problems may deem it expedient to pay bonuses to employees as an investment in goodwill, or to make office improvements with the thought in mind that the government is contributing a portion of these expenditures. OVERHEAD AND TOTAL COSTS So far the discussion of overhead costs has been cast in terms of dollars and cents. When the ratio of overhead to total costs is computed, one interesting feature is presented. Table 15 shows the fluctuations of this ratio, in percentage figures and in relatives adjusted for seasonal variation, on the same base as the other series, 1920-23. The course of these fluctuations differs somewhat from those of overhead expressed in dollars. The proportion of total costs represented by overhead is like overhead alone in rising during 1920, but unlike overhead alone in that it rose and fell irregularly from 1921 on, with a slight tendency to decline. This means that the sum of the other two elements of total costs constituted a great proportion of the total after 1921 than before. The problem of overhead costs is the problem of idle capacity.3 It would be extremely interesting to compare overhead costs with some measure of the capacity at which these plants were producing at various periods of the cycle. If such a comparison were possible, it is likely that the lump sum of overhead would be too rough for nice work, and would 'See J. M. Clark, "The Economics of Overhead Costs," p. 1 and passim. OVERHEAD COSTS 47 - need to be broken into its constituent elements before any precision could be attained. Unfortunately no measure of productive capacity, used and unused, is available. TABLE 154 FLUCTUATIONS IN THE RATIO OF OVERHEAD COSTS TO TOTAL COSTS, 13 FIRMS. 1920 1921 Percent Adjusted Percent Adjusted actual relatives actual relatives January.............. 20.0 77.2 27.2 105.0 February............... 21.0 75.0 30.0 107.1 March.................. 18.9 74.4 28.8 113.3 April................... 22.2 82.2 29.2 108.1 May.................. 20.9 80.7 29.9 115.4 June................... 19.9 78.3 28.5 112.2 July................... 20.8 78.5 29.6 111.7 August............... 20.0 78.7 29.1 114.6 September.............. 22.4 84.5 29.1 109.6 OCtober.............. 22.6 93.8 27.0 112.0 November............. 26.6 98.5 27.5 101.8 December.............. 31.0 108.8 30.9 108.4 1922 1923 Percent Adjusted Percent Adjusted actual relatives actual relatives January............... 29.1 112.4 25.6 98.8 Pebruary.......... 30.7 109.6 28.5 101.8 March................ 27.2 107. 1 25.9 102.0 April................... 27.0 100.0 28.8 106.6 May................... 26.9 103.9 25.4 98.1 June................... 26.5 104.3 29.8 117.3 July...............:111..... 29.1 109.8 30.7 115.9 August.......... 28.2 111.0 27.8 109.4 September............. 30.0 113.2 28.4 107.2 October................ 27.1 112.4 25.0 103.7 November............. 26.7 98.9 35.3 130.7 December.............. 27.9 97.9 25.0 87.7 'See also Chart 4. 48 THE NEW YORK PRINTING INDUSTRY CHART 4 FLUCTUATIONS OF PRESS IMPRESSIONS AND OF THE RATIO OF OVERHEAD COSTS TO TOTAL COSTS, 1920-1923, AD, JUSTED FOR SEASONAL VARIATION. Average 1920-1923= 100. Unit, one percent. I 35 zo~~ 1 L I I III II Prr i 1 1 1 1.____-. _,..-. _A.n. 12~~~~~~~~~ SS o,~~~~~o ____ _ _- -]r X Ivocu /*ZJ WI?} CHAPTER VI FLUCTUATIONS OF TOTAL COSTS AND PROFITS TOTAL COSTS The fluctuations of the total cost series are found in Tables 7 and 8, and in Chart 3. Since total cost is ascertained by adding payroll, purchases, and overhead, the fluctuations of this series might be expected to amount to nothing more than a compromise between the differences of its constituents, unless one of them were more influential than the other two. So it turns out. The peak of the series was in June 1920, delayed to that point by the force of the purchase peak assisted by a minor peak in overhead. The trough of the total costs came in May 1921, when all three contributing series happened to decline together. The relatives, adjusted for seasonal variation, are always somewhere between the relatives for the other series. The indexes of seasonal variation are sometimes inside, sometimes outside, of the compass of the payroll and the overhead indexes, and the outside cases must be ascribed to the influence of purchases, for which no seasonal index could be computed. Total costs fluctuated more widely than payroll or overhead, but less violently than purchases. It is interesting to note how the composition of total costs varied from year to year. Table 16 below shows these variations. Payroll and overhead vary together, in this showing, at the expense of purchases. The payroll and overhead proportions rose in 1921 over 1920, overhead rising more than payroll, and 49 50 THE NEW YORK PRINTING INDUSTRY TABLE 16 VARIATION IN THE PROPORTION OF TOTAL COSTS REPRESENTED BY PAYROLL, PURCHASES, AND OVERHEAD, 1920-1923. Percentage Percentage Percentage Year payroll to purchases to overhead to Total cost total cost total cost total cost 1920 44.3 33.7 22.0 100.0 1921 47.8 23.4 28.8 100.0 1922 46.1 25.9 28.0 100.0 1923 46.6 25.6 27.8 100.0 1920-23 46.1 27.5 26.4 100.0 purchases fell off. As overhead and payroll slowly fell off, purchases made moderate gains. At the end of the four-year period, the relation which had existed in 1920 had not returned, for payroll and overhead showed a net gain and purchases a net loss, with respect not only to the 1920 level, but also to the level of the four-year average. In Table 14 the payroll for 1922 represented only 32.5 per cent of the total costs. This was the proportion which held for Class G establishments in the U. T. A. survey. The higher figure of 46.1 for the same year, for the New York establishments, may be explained partly by noting that a large proportion of the work of these plants is the manufacture of periodicals. On all large publication printing the publisher furnishes the printer with practically all material to be used in the printing of his magazine, including paper. This, and the higher wage scale of New York City, would account for the difference. PROFITS The chief significance of the total cost series lies not in itself, but in its relation to sales (billings). The difference between them is profit or loss, according as sales are greater than costs or less. This difference is reported in Tables 7 and 8.1 See also:harts 5 and 6. TOTAL COSTS AND PROFITS 51 CHART 5 FLUCTUATIONS OF PRESS IMPRESSIONS (ADJUSTED) AND PROFITS (UNADJUSTED), 1920-1923. Average 1920-1923 = 100 Unit, one percent. rua ~ ~ r -l l- l l \A l l l l l: l _ A32~t~~~~ t tt. t /TJ t fLeo _>_ HX i V__X__ -zoo EI_ < I_ - 85 * ___-A__.. i.s TLr How did the dollar profits fluctuate? Like the other series (except employment, purchases, and overhead) the peak was reached in March 1920. It was a high peak, for no other month approached it by a closer margin than $57,000. The trend downward began abruptly, halted with the seasonal influence of the year-end rush, then continued down-grade until the December business of 1921 retarded the decline. From that point the decline was irregular; January 1922 showed the first loss, another was registered in March, and from June until November no profits were recorded, showing a net loss for the year. The trough, or greatest loss, was taken in October 1922. Three months with losses were experienced in 1923, but the trend by this time was distinctly favorable. November 1923 was a freak month, showing a profit well over two-thirds of the 1920 peak. This was 52 THE NEW YORK PRINTING INDUSTRY probably a reflection of the rush of production noted in October. In intensity, the range was from a profit of $380,434, to a loss of $71,788, a decline of 119 percent2 of the peak profit. CHART 6 FLUCTUATIONS OF PRESS IMPRESSIONS AND PROFITS, 1920 -1923, ADJUSTED FOR SEASONAL VARIATION. Average 1920 -1923 = 100. Unit, one percent. 130 1l'8 tti t 1 I~ I T 1 1I 1 I 1 1 1 1. I 1 1 1 1,2 RT I I T T I I T 1 1 1- I I I T I 1 I Ii-.,__AA_ ___ ---L v -.. 4 S Y t _ _ _,- ____ I.. 7o~ SoC| t T i t t; 0 /320 ag1 1922 mu2. How does the general trend of profits compare with the general trend of production? The answer to this question is really the answer to the main problem of the investigation, namely, what, at any point of time, is the economic condition of the industry? Taking the adjusted relatives of production, and comparing them with the unadjusted profit and loss items in dollars, it is seen that during the first year profits fell away faster than production, recovered at the beginning of the second year and then fell away more slowly than production. The trough in profits was not reached until a year after production had entered the revival stage. During the third year, 'A decline of more than 100 percent is possible when negative items (in this case the losses) are included in the series. TOTAL COSTS AND. PROFITS 53 production was slowly reviving, while profits were irregularly declining. During the fourth year, a relapse in production prevented a rise over the four-year average except during four months, and profits, although on the mend in general, rose and fell irregularly and registered three losses. A distinction may be drawn between the short-period and long-period correspondence between the two series. At particular points of the four years, profits tended to vary directly with the production. For example, the January 1921 increase in production was accompanied by an increase in profits, and the fall in production in the following month brought a decrease in profits as well. But in general, the profit series was unable to rise when production's trend turned definitely upward; profits continued to fall, with only sporadic recoveries, for twelve months more; This is the most important single conclusion to be drawn from the investigation. The meaning is apparently that although within short intervals profits depend upon production, and an increase in production means that profits increase also, nevertheless in the long run more fundamental influences are at work to destroy the close relationship. The most important of these is probably the business or financial policy of the establishments here represented. Prices do not seem to decline with production, but after production has declined. Prices are reduced not in anticipation of a reduction of business, but after the business has gone; the barn is locked after the horse has been stolen. Yet the owners and managers of these establishments do not have entire control of the situation. A large part of their work is the manufacture of periodicals, and it is a trade custom to contract in advance for such work. Moreover, it is not a trade custom for printers to ask their customers to interrupt contracts, revise them downward in price and extend the lower price level to a date later than the normal expiration of the contract. 54 THE NEW YORK PRINTING INDUSTRY But if these data may be taken as representative of the New York centre, and this depression as typical of other depressions, some such interruption is indicated. Some evidence in support of the suggestion is had from the comparison of large and small firms made on page 16 and page 20. In this comparison it was shown that the large firms did not recover as quickly, in 1921 and 1922, as small firms. When this fact is viewed in the light of the greater predominance of contract work in large than in small establishments, it points to the contract system as a prime cause of the delay in price adjustment and the probable necessity of making more drastic cuts when contracts finally expire.3 'This suggestion is made tentatively, for the evidence on which it rests is too slender to support a categorical recommendation. Yet it indicates that the Research Bureau of the United Typothetae of America might arrive at valuable results by studying the problem with the aid of the voluminous and complete cost data in its possession. CHAPTER VII PRINTING AND GENERAL BUSINESS CONDITIONS As a final step it is worth while to see how the book and job branch of the printing industry in New York City fluctuates in comparison with other aspects of printing, and with general business conditions. Such an inquiry will accomplish- two objects. It will serve to give a modicum of precision to the relation between this branch of the industry and the many measures of general business conditions which have been used in printing arbitrations in the past." There are too many of these measures to examine them all, but if they can be understood as represented by an index of general business conditions and the relation of this index to the printing industry demonstrated, it will be possible for the parties to future arbitration Proceedings to avoid the pitfalls of haziness and vagueness which have beset them in the past. The second object which the analysis should accomplish is to offer a broad basis for the prediction of printing fluctuations. An infallible formula, in the present state of statistical wisdom and with only four years of data to go on, is of course impossible. But if.these series can be related to some published index which is reliable, and which is to be continued, it will be possible for the industry to carry its own figures farther and continue the comparison in the future. Out of such a comparison a reasonably close forecasting method may develop. PRINTING AND ELECTROTYPE SALES Two sets of figures are available in fields related closely to printing. The first is the electrotypers' sales. This series, confined to New York City, was developed by Mr. C. A. Hok'See above, pp. 2 and 3. 55 56 THE NEW YORK PRINTING INDUSTRY inson,2 and is derived in the same manner as the printing series. Fifteen electrotyping establishments are represented. Since prices changed so little between 1920 and 1923 in electrotyping companies, the index is practically a physical one. It is closely related to printing, because printers consume the entire product of the electrotyper; and a large part of that product goes to the book and job branch. TABLE 17 3 ADJUSTED RELATIVES OF ELECTROTYPE SALES, 1920-1923. (Average of 1920-23 = 100) 1920 1921 1922 1923 January................ 117.6 102.7 87.7 97.1 February............. 109.1 101.9 91.2 97.8 March.................. 119.1 89.9 93.2 96.3 April................... 130.0 88.7 86.7 95.6 May.................. 125.4 85.5 95.7 96.0 June.................. 125.8 87.8 91.8 98.0 July.................... 119.9 90.4 90.8 98.5 August................ 120.2 94.8 92.1 98.5 September............ 119.5 95.6 91.9 98.7 October..1.......... 104.9 92.7 96.6 100.4 November.............. 103.4 94.5 97.2 99.4 December............... 96.2 99.2 95.7 100.6 The peak of this series came in April 1920, one month after the production peak in printing. It showed a decline as irregular as that in printing, but slower, and it reached its trough in May 1921, five months before printing production hit bottom. It recovered well in the fall of 1921, and after a relapse in the beginning of 1922 it revived gradually and evenly until the end of the period, barely reaching the average line of the four years at the close. The main difference between electrotyping and printing appears in the different peak and trough dates, and in the evenness of the revival. Also, electrotyping did not decline as far as printing. In the minor irregular swings printing seems to anticipate electrotyping more often than not, by about a month. 'On the basis of figures furnished by the Employing Electrotypers' Association. 'See also Chart 7. PRINTING AND GENERAL BUSINESS 57 CHART 7 FLUCTUATIONS OF PRESS IMPRESSIONS AND OF THE SALES OF 15 ELECTROTYPE FOUNDRIES (PRICES CONSTANT), 1920-1923, ADJUSTED FOR SEASONAL VARIATION. Average 1920-1923 = 100. Unit, one percent. 130 IC ~ _. - I _ \ _9_1__ _ _Xi;:- -- - v B r ---- x lwG ~ 1 /911 1 PRINTING AND MAGAZINE ADVERTISING The second series allied to printing is the total of advertising in magazines. The figures are taken from the Harvard Economic Service4 and are based on the average of 1919-1921 as 100.6 The figures are corrected for seasonal variation. Chart 8 shows the comparison. This series compares very closely with printing. The 1920 peak came at the same time, as well as the 1921 trough. The advertising curve declined gradually in 1920, and then, with the printing curve, dropped sharply in the early months of 1921. But advertising dropped farther than printing. Its revival proceeded gradually but at a greater rate than printing, and its 1923 peak followed the printing peak by a month. "'Advertising and the Business Cycle," Weekly Letter No. 9, Vol. III, March 1, 1924. 'See above, page 25. 58 THE NEW YORK PRINTING INDUSTRY Magazine advertising reached a slightly higher level by the end of the period than printing, but neither of these series exceeded the average line (of 1919-21 and 1920-21, respectively) by any considerable amount. CHART 8 FLUCTUATIONS OF PRESS IMPRESSIONS AND TOTAL ADVERTISING IN MAGAZINES, 1920-1923, ADJUSTED FOR SEASONAL VARIATION. Average 1920-1921 = 100. Unit, one percent. _l I I _ _ I -1 X__-4 _44 _ ___ _ t — and magazine advertising during the period 1919-23, the Harvard Economic Service6 makes several comments which help in explaining the failure of printing production to revive much beyond its four-year average. It says: "Comparison between the curves for newspapers and for magazines shows that the latter curve has a much greater range of fluctuation than the former. This condition is due partly to the abnormally high level of magazine advertising in the years 1919 and 1920, OP. cis., pp. 47-49, PRINTING AND GENERAL BUSINESS 59 but this fact is insufficient to account for all of the dissimilarity. An additional reason is found in the tendency for magazine advertising to be used chiefly by producers and distributors on a national scale, whereas the retail distributor dominates in newspaper advertising. Consequently, magazine advertising may be expected to fluctuate more widely than newspaper advertising, just as wholesale prices fluctuate more widely than retail prices. "... The unusually large volume of magazine advertising in 1919 and 1920 would tend to lower the index for magazines for the later years. the figures for 1919 and 1920 are affected so sharply by the heavy advertising expenditures encouraged by excessive federal taxation,." Partly because of excessive federal taxation, then, magazine advertising was unduly inflated in 1919 and 1920, and since the base-line in this series is affected by the average of these two years and the next year, the average is high. The wide fluctuation of magazine advertising is explained by the national aspect of periodical advertising. This assists in explaining the fluctuations of printing production, for a considerable portion of the work of the book and job branch of the printing industry is the manufacture of periodicals, and the size of the periodicals (and hence the work of making them) depends to some extent on the quantity of advertising which they carry. The chief conclusion to be drawn from these two comparisons is a negative one. Electrotype sales vary closely with printing, for what the electrotypers do depends on what the printers order. Magazine advertising varies closely with printing, for what the printers do depends, to some extent at least, on the optimism of those who advertise on a national scale. But although nothing was found in these comparisons which is difficult to explain, some precision has been given to the relations between the printing production series and the two others. In subsequent arbitrations it will be possible to check so THE NEW YORK PRINTING INDUSTRY the printing fluctuations against those in advertising and electrotype sales, with the assurance that if all three show a determined trend in the same direction, the trend is more significant than if it were found in only one. There are three series outside the printing industry which will serve to throw some light on the relation of printing to business in general.- They are, first, the employment figures of the New York State Industrial Commission;7 second, the volume of manufacture, (all lines combined) published by the Harvard Economic Service;8 and, third, the Harvard index of general business conditions, known as the ''B-curve."9 EMPLOYMENT IN NEW YORE STATE The New York State employment figures were recomputed so that they would have the average of 1920-23 as the baseline. The series, so calculated, compares with employment in the industry under discussion as follows: TABLE 1810 EMPLOYMENT IN N. Y. CITY BOOK AND JOB PRINTING AND IN N. Y. STATE FACTORIES, 1920-23. 1920 1921 1922 1923 City B State City B State City B State City B State and J factoriesand J factories and J factories and J factories Jan. 125 117 99 89 93 88 96 104 eb. 137 116 94 90 99 91 98 105 Mar. 131 119 89 91 98 92 98 108 April 120 118 89 90 98 91 101 107 May 116 116 89 88 100 92 107 -106 June 117 115 93 86 97 93 101 105 uly 113 115 90 84 98 93 97 105 Aug. 116 113 93 84 97 95 88 104 Sept. 106 112 90 88 101 97 90 104 Oct. 99 110 91 90 95 100 102 105 Nov. 96 103 94 90 97- 102 99 104 Dec. 97 95 91 90 97 104 94 102 7Standard Daily Trade Service, Statistical Bulletin No. 10, October 20, 1923, page 154. 'Review of Economic Statistics, Prel. Vol. V, June. 1923, Supplement 1, p. 163 and Vol. VI, July, 1924, Supplement i, p. 128. 'Ibid, Prel. Vol. V, June, 1923, Supplement 1, p. 167, and Vol. VI, July, 1924, Supplement 1, p. 174. "See also Chart 9. PRINTING AND GENERAL BUSINESS 61 CHART 9 FLUCTUATIONS OF EMPLOYMENT IN PRINTING AND IN NEW YORK STATE FACTORIES, 1920-1923, ADJUSTED FOR SEASONAL VARIATION. Average 1920-1923 = 100. Unit, one percent. Z 'I L - I I I ~~InI11ItIII i /30~~~~~~~~~~~~~~ I, tO, /I21 '~~.,:zIi I _____ Copn _ ________.o A! *enl J.Comparing these two series, it is seen that the peak in both cases was the same, i. e., in March 1920. The New York State figures declined more gradually than printing employment and reached the trough about three months later. Employment touched bottom in May 1921 in printing, and in the general manufacturing 'of the state in July and August. The revival was at about the same rate, but in October 1922 the state factories outran printing and finished on a higher level than printing. It may be suggested that the higher start of printing employment accounts for its poorer finish, as in the. case of production. As in the previous comparisons, the factory series does no more than reinforce the conclusions which may be drawn from printing employment alone. PRINTING AND GENERAL MANUFACTURING A more interesting comparison is that between printing production and the Harvard figures for the volume of manufacture (all lines combined)." Recomputing the latter index "See Chart 10. 62 THE NEW YORK PRINTING INDUSTRY CHART 10 FLUCTUATIONS OF PRESS IMPRESSIONS AND OF THE VOLUME OF MANUFACTURE, ALL LINES COMBINED (HARVARD), 1920-1923, ADJUSTED FOR SEASONAL VARIATION. Average 1920-1923 = 10C0. Unit, one percent. e15~M, LS i 77 1jEET II _IJI - -- II _ / _~ _ _ — I ' 1111 oo-_-____I.,I 8Q~~- _- -7 l O -i_ r - - - - - 70 _ _~~ _ __ _- 3- - '901 /921 /912! /92J on the 1920-23 base, it is found that the peak of all manufacturing came in January 1920, two months earlier than the peak of printing production. The total manufacturing peak was not nearly as high, the relatives (on the 1920-23 base) being 119 for total manufacturing and 147 for printing. The decline in total manufacturing was rapid during 1920, but not as steep as that in printing. Total manufacturing touched bottom in May 1921, five months ahead of the printing trough. The recovery in total manufacturing proceeded at a faster rate than in printing, and after April 1922 the printing curve was never above the total manufacturing curve.. Total manufacturing reached the 1923 peak in March and May, five and seven months ahead of printing. If the month of October 1923 is considered a freak month in printing, and the next highest 1923 month, November, be considered also freakish, the print PRINTING AND GENERAL BUSINESS 63 ing peak may be said to have come in January and May 1923. In the more important details of the fluctuations the series seem to run together, with no lag in timing. But in the general swings, total manufacturing appears to precede printing production by a period of from two to five months. If this relationship is maintained the volume of manufacturing can be used as a rough predictive index for printing. But it cannot fairly be used in this manner without further comparisons which show that the correspondence is an economic matter, not an accidental one PRINTING AND GENERAL BUSINESS Lastly, how does the printing business compare with general business conditions? Of the many indexes of general business conditions which are at hand, the Harvard Economic Service index has been selected for this comparison. This selection was largely arbitrary, but it was based on the current wide acceptance of the scientific character of the service, and in part also on its promptness and availability. The second curve was selected, of the three which are published under the heading "Index of General Business Conditions." This is the "B-curve," which is composed of bank debits outside New York City, and commodity prices. The second of these ingredients is a price index, the first is the product of trade and prices. The corresponding index in printing is billings, which is a function both of production and prices. But the influence of prices, which affect the "B-curve" twice, is felt in billings but once. A comparison of these two indexes'2 shows that the peak of the Harvard "B-curve" came in July 1920, while billings reached a high point in March. The actual trough of billings (not the adjusted trough, which we have regarded above'3 as misleading) was registered in July 1922, six months after the general business low point. The printing revival ended simul13See Chart 11. uSee page 35. 64 THE NEW YORK PRINTING INDUSTRY CHART 11 FLUCTUATIONS OF (A) PRINTING BILLINGS AND OF (B) GENERAL BUSINESS CONDITIONS (HARVARD "B-CURVE"), 1920-1923 ADJUSTED FOR SEASONAL VARIATION. (a) Average 1920-1923 = 100. Unit, one percent. (b) Computed normal =0. Unit, one standard deviation. 130;so3111 1 1 1 1 1 1 1 1 1 11 1 1 1 1 1 1 1 1 + /00 O~~~~~~~~~~~~~~~~ '920 WIt /92! ft taneously with the general revival, in May 1923. The printing curve is much less regular than the "B-curve," because it concerns only one industrial group in a single city, and because the inclusion of a wide variety of industrial and commercial classes, in all districts, which is implied in the index of general business conditions, would naturally produce a smoother surface. Billings did not reach as high a level in 1923 as the "B-curve," probably for the reason noted above, namely, that the former series started high. But on this point there is less certainty than in the other cases, for it is not possible to compare the violence of the fluctuations of two series unless they are cast in the same terms. The Harvardbcurve is expressed in units of the standard deviation of the percentage deviations PRINTING AND GENERAL BUSINESS 65 from the line of secular trend, whereas the billings are expressed as relatives. In order to compare billings with a series composed, like itself, of trade volume and prices, and cast in comparable terms, the index of bank debits outside New York City was examined. Chart 12 shows the comparison. In this chart both series are adjusted for seasonal variation. The base of the debits series is the average of 1920-21; since this base differed from the average of 1920-23 by only one-half of one percent, it was not shifted. CHART 12 FLUCTUATIONS OF (A) PRINTING BILLINGS AND OF (B) BANK DEBITS OUTSIDE OF NEW YORK CITY, 1920-1923, ADJUSTED FOR SEASONAL VARIATION. (a) Average 1920 -1923 = 100. Unit, one percent. (b) Average 1920-1921 =100. Unit, one percent. proceeded at almost the same pace. Stable bottom was reached a_ _- _-~_-X R, The high point in outside bank debits came in January 1920, two months before the billings peak. The decline in each series proceeded at almost the same pace. Stable bottom was reached in outside bank debits in June and July of 1921, a year ahead of the actual billings trough. The debits revived slowly, reach, 66 THE NEW YORK PRINTING INDUSTRY ing a new peak in May 1923, simultaneously with the peak in billings. Are these two series, the Harvard "B-curve" and the index of bank debits outside New York City, of any use in predicting the printing fluctuations? The answer must be no, for no such constant relationship appears to exist between these series and billings as was found between general manufacturing and production. In the first comparison, the 1920 printing peak preceded the general peak, the printing trough followed the general trough, and the 1923 peaks coincided. In the second comparison, the general peak preceded the printing high point by two months, the general trough ran twelve months ahead of the low point in printing, and again the 1923 peaks coincided. This is not a hopeful finding. In the long run one may expect the printing series and the general series to run together, largely because the printing industry, as one printer said in the course of an arbitration hearing, is "not carried on in an economic vacuum." But for shorter runs neither general index is useful in establishing predictive norms. Thus the two-fold task set at the beginning of this chapter has been accomplished. The many measures of the business cycle which have been presented to past arbitrators without precision, or for that matter with little discrimination, have been tacitly summarized in two indexes of general business conditions, and these indexes have been shown to have no consistent relation to the printing business except in a very general way. The relation of some aspects of printing to electrotypers' sales and magazine advertisements has been demonstrated. And lastly, out of these comparisons one series was found which seems to promise, provided further study is made, to give a basis for forecasting the trend bf printing production. This forecasting must be based on two assumptions. The first is that the relation which has held in the past between printing production and general manufacturing will continue PRINTING AND GENERAL BUSINESS 67 to hold in the future; and this is only a probability. The second is that the composition and business policy of the New York book and job printers will remain as they are; which is likewise only a probability. Accordingly, any predictions on the basis of these figures must be managed with great caution and advanced with little assurance. CHAPTER VIII THE ECONOMIC CONDITION OF THE INDUSTRY This study may now be summarized in a few paragraphs. It arose out of the apparent need for a more thorough knowledge of the economic condition of the book and job branch of the printing industry in New York City, for use in wage negotiations, than has existed hitherto. The arbitration proceedings between the several printing trades unions and the Printers' League Branch of the New York Employing Printers' Association had contained many references to the "economic condition of the industry" without giving specific attributes to that condition. Measures of this condition were found in various fluctuations of the industry during the period beginning in January 1920 and ending in December 1923. The fluctuations of production (press impressions), employment, sales, costs, and profits were noted. These data were reported by seventeen large concerns (not all of whom reported all the information). This number represented one tenth (in volume of business) of the total book and job industry in New York, and one fourth (in volume of business) of the membership of the N. Y. Employing Printers' Association. A study of the reporting firms in comparison with 40 smaller ones did not show any reason for discarding the data as insufficiently representative. Seasonal variation was eliminated where possible by rough methods which would have been inadequate for figures extending over a longer period, but which were considered as accurate as possible under the circumstances. The series 68 THE ECONOMIC CONDITION OF THE INDUSTRY 69 were presented as relatives on the base line of the average of the whole period 1920-23 as 100. The timing and intensity of the fluctuations of the various measures were compared. The data were found to be internally consistent, so that conclusions drawn from them have a logical validity. The comparisons showed that: 1. In general, the peak of production, employment, sales, costs (except overhead), and profits came in the early part of 1920, and was a high one; the decline was rapid; with the exception of sales and profits the trough was reached in the fall of 1921; the revival was gradual through 1922 and 1923, and the peak reached in 1923 was barely above the base line, the average of the four years. 2. Sales showed one notable divergence from the other series; they did not reach the trough until the summer of 1922. The explanation was advanced that these large concerns were restrained from early and drastic price-cutting by the widespread use of annual contracts in this class of printing. 3. Profits also showed peculiar behavior. The trough was not reached until October 1922, and profits were losses throughout 1922 and in three months of 1923. But production had been reviving for a year before the greatest loss was registered. Profits, then, do not vary simultaneously with production, but in a depression (if this depression was typical), they continue to fall for some time after production has been rising. One might expect profits and production to rise and fall together. The responsibility for the lag was laid partly at the door of the contract scheme, which delays adjustments, and partly to the highly competitive nature of the industry. 4. Labor suffered more from unemployment than from wage reductions. The plan of some employers for managing the unemployment problem runs contrary to this finding, for it is often felt that labor is better off in a depression if it will accept the reduction necessary to keep it at work. The cause 70 THE NEW YORK PRINTING INDUSTRY of the condition was ascribed to the rigid character of wage agreements between employers and unions. 5. Overhead costs, in the somewhat peculiar sense in which the word was used here, rose rapidly in depression, not only relatively, which might be expected, but absolutely, which might not be guessed so easily. The explanation advanced was the probable increase in sales effort during 1920 and the first few months of 1921. Finally, the close relation of the printing series to a series showing electrotype sales and to the volume of magazine advertising was demonstrated. Comparisons were madewith the figures for employment in New York State factories, the volume of manufacture (all lines combined), and with an index of general business conditions. Of these, only the curve of the volume of manufacture was found to be of any service in predicting the course of printing fluctuations. This index, although it needs to be studied farther to see whether the relation noted continues to hold, seems to be useful in predicting the fluctuations of printing production from two to five months ahead. WAGE ARBITRATIONS How can these materials be used in wage arbitrations? Two relations, among those noted above, stand out as being of particular use to arbitrators-of wage difficulties. They are, first, the relation of the Harvard curve of manufacture, all lines combined, to printing production, and second, the relation of the curve of printing production to the rise and fall of profits. These two relations have been plotted on Chart 13. As previously noted, the fluctuations in total volume of manufacture appear to anticipate the fluctuations in printing production by an interval of from two to five months. It is important again to emphasize the need for continuing study of this relationship, for its reliability has not yet been proved. THE ECONOMIC CONDITION OF THE INDUSTRY 71 Printing production, in its turn, (at least when it is near the trough) seems to forecast losses instead of profits about a year ahead. CHART 13 FLUCTUATIONS OF PRESS IMPRESSIONS, PROFITS, AND VOLUME OF MANUFACTURE, ALL LINES COMBINED (HARVARD), 1920-1923, ADJUSTED FOR SEASONAL VARIATION. Average 1920-1923=100. Unit, one percent. IN INa r ~~~~~~~~~~~~* II I.,. M. - - f rt ~_ _ _ _ _ _ -> _ 7; _ _ _ _ _ _ - - - e-' * - - __ ____.- _ — - - - II -:u 'su 0 2tasu a..~ - X-_,eno /92i/92 By removing the middle term (printing production) it is possible to argue that a decline in the total volume of manufacture considerably below the normal line may be used to predict probable losses in the industry fourteen to seventeen months hence. These considerations can be used by the arbitrator, or by the disputants, in settling on a wage policy. CONCLUSION This study started with the hope of finding some "definite, authoritative, and conclusive". measure-of the economic con 72 THE NEW YORK PRINTING INDUSTRY dition of the printing industry. How far has that hope been realized? The economic condition of the industry, in the sense that those words contain here, is the ability of the industry to pay wages in greater or less amounts. Is it now possible to say at a given time that the industry can or cannot afford a wage increase, or that at a given moment the wage burden of the industry must be reduced to save the industry? Has the ability of the industry to pay wages been measured? To a certain extent that ability has not been measured. The unions concerned in the printing arbitrations would presumably agree to the statement that their employers are entitled to a reasonable return on their capital investment. The rub is with the word reasonable. For want of a suitable definition the word some may be substituted; the employers are entitled to some return. Obviously, when there is no profit at an existing level of wages there is no ability to pay an increase. But when there is a profit, no way exists of determining how much of it is reasonable and how much of it ought to be devoted to wage increases by virtue of being "unreasonable." This is purely a matter of opinion, and as such it must eventually be thrashed out in future arbitrations. The probable basis of the dispute will be not what percentage of costs may be taken as "just profit," but what percentage of the investment. As the parties warm to the dispute, the problem of justice in accounting will need attention. Fortunately the United Typothetae of America has been active in devising a sound system of accounting for printers all over the United States, and this system is coming rapidly into wide-spread use. The defense of this system will be much easier because of its uniformity and because of the attitude which prompted its designers, namely the welfare of the industry as a whole. Presumably it is a conservative system, and the burden of proof will lie with those who propose changes in it. If the study has failed to establish a definite measure of the THE ECONOMIC CONDITION OF THE INDUSTRY 73 ability of the industry to pay more wages or less wages, it is thus because that condition is not altogether a matter of fact, but rather a matter of attitude and opinion. To a certain extent, the economic condition of the industry has not been measured because it is not a measurable condition, having no concrete dimensions. Yet in some degree this study has contributed information which should make the discussion of opinion and attitude more intelligent than it would otherwise be. It has provided a statement of the relation of production and profits; it has pointed out the part of the business cycle in which profits disappear, and with them the ability of the industry to carry more than its current wage burden; it has shown how to tell when this part of the business cycle is approaching; in short, it has provided enough information so that the dispute about just profits, when that issue is eventually raised, may be based in some measure on matters of fact and not confined wholly to guessing. Obviously, the factual basis has not been attained in anything like the degree which will become possible as the analysis is continued. I, INDEX Accounting surveys, 3-4, 14 n. Advertising, 57-59, 70. Arbitration, v, 55, 59, 68, 70-72. Bank debits outside N. Y. City, 63, 64. Base, choice of, 25, 69. "B-curve," Harvard, 63-64, 66. Billings, 9, 15-16, 18, 20, 31-40, 50, 63-66, 68-69. Bindery Women's Union No. 43, 6, - 39 n. Book and Job Printing, v, 7. Bureau of Labor Statistics, v. Capacity, 46. Clark, J. M., 46. Clay, H., 39 n. Contracts in printing, 20, 40, 53-54, 69; Census of Manufactures, 7, 12. Costs, overhead, 10, 41-50, 68, 70. Costs, total, 10, 33, 42-44, 46-50, 68. Data, 8 ff., 28-29. Reliability of, 20-21. Representative character of, 11 ff. Statistical treatment of, 22-27. Economic condition of the industry, v, 2, 68-73. Electrotype sales, 55-56, 59, 70. Employment, 9, 31-40, 60-61, 68-69. In N. Y. State factories, 60-61,70. Firms, cooperating, 11, 15. Firms, large and small, 13-15, 54. Fluctuations, intensity of, 36-38,69. timing of, 32-36, 69. Forecasting, 55, 63, 66-67, 70-72, 73. General business conditions, 55-66, 70. Harvard Economic Service, 25-26, 57-58, 60-61, 64, 70. Hokinson, C. A., vi, 27, 39 n., 46, 55. Hours, productive or chargeable, 9. Hours worked, 9, 37, 38. Impressions (see Press impressions). Job Pressmen and Job Feeders Union No. 1, 6, 39 n. Kendrick, E. A., 3. King, W. I., 14 n., 20 n. Labor efficiency, 36-37, 39-40. Losses (see Profit and loss). Macaulay, F. R., vi, 26 n. Mailers Union No. 6, 6, 39 n. Manufacture, volume of, 61-62, 66, 70. Method, 4-5. Mitchell, W. C., vi. N. Y. Employing Printers' Association, v, 1, 42 n., 68. N. Y. State Industrial Commission, 60. Overhead (see Costs, overhead). Paper Cutters Union No. 119, 6, 39 n. Paper Handlers Union No. 1, 6, 39 n. Payroll, mechanical, 10, 15, 18, 31-40, 49-50. Persons, W. M., 26 n. Premium men, 3, 35. Press impressions, 8, 22-24, 31-40, 44, 52-53, 56-57, 59, 61-62, 66, 68-69, 71, 73. Presses, types of, 8, 23-24. Prices, 36-37, 54. 75 Is.., * *, C. 0. I 3 c 76 IN] Printers' League, 1-2, 6-7, 68. Printing Press Feeders and Assistants Union No. 23, 6, 39 n. Printing Pressmens Union No. 51, 6, 39 n. Production (see Press impressions). Profit and loss, 10, 27, 50-54, 68-69, 71-73. Purchases, 10, 27, 32-33, 37, 46, 49-50. Revuiew of Economic Statistics, 60 n. Rorty, M. C., 20 n. Rosett, M., 3. Sales (see'Bilings). Scope, 6-7. Seager, H. R., vi. DEX Seasonal variation, 25, 28, 31, 44-45, 68. Sheet Straighteners Union No. 119, 6, 39 n. Silcox, F. A., vi. Standard Daily Trade Service, 60 n. Taxes, 46, 59. Total costs (see Costs, total). Typographical Union No. 6, 2, 6, 39. Unemployment, 2, 38, 69. Unions, v, 38-39, 68. United Typothetae of America, 9, 11, 14 n., 41, 43, 54 n., 72. Wage agreements, 70. Wage burden, 39, 73. A I I rI i I I t J I f ~. t -(-.: -- -~5' i `;7 Ji~; ~i. —: - —. -ti i~ t~?_j -t. i '~ —;i i; I,: i — I i -~I-t.~:-;i —: ~- I. -i:-i I!.i i_,;; I ~ —~ r i-:j -~ -- i;i; Y1 -.L,=.- i t -i / --- c -7 i'/ x. i ii C I ! l0LJB R 4966. ~ ~ i. i.. \e> 2'.;.: 0 - i-. < 31 U. > 1z. z X -. -r I,. 'K' '' *: J 1 _..~~~~~~~..!~;, ~N.' a.':~ ':;~,' 'S S 0, I~r A ^ I ~; i ~,I:.*.x:, ',. I.^... h~~~~~~~. I:rt N,D S;.'\'',r., I:f ~i~\:..' C,,.'. tA..;. 1.. " '''f"'-''iN ''-.i' '.i'n2 UNIVERSITY OF MICHIGAN: I1 Clt ill I 1 11 11 11 111 11 I1 I 3 9015 03189 2519....... - v-. —1. — rt. -.. I Craig t Of the Printi Inutry i ey o..... 014075 -. /. I ',., "j,.i; I S 1. t -,,g..,^ I. " '0 X 0 ''S7,"i.. i 4'-'"-.j; '''V ~i~S;- 9Asr' t~i~;;i'~/ 9,:I.*L:i r' r \'~~ i!. LI ' ~1~-i iyr ?ii,, i~, ~;. tt ', i'L':: >'! s,i, XI- l'<..f st