chapter 4 / the american baby boom in historical

This PDF is a selection from an out-of-print volume from the National Bureauof Economic Research

Volume Title: Population, Labor Force, and Long Swings in Economic Growth:The American Experience

Volume Author/Editor: Richard A. Easterlin

Volume Publisher: NBER

Volume ISBN: 0-87014-474-X

Volume URL: http://www.nber.org/books/east68-1

Publication Date: 1968

Chapter Title: The American Baby Boom in Historical Perspective

Chapter Author: Richard A. Easterlin

Chapter URL: http://www.nber.org/chapters/c1119

Chapter pages in book: (p. 77 - 110)

F

NGS

)rrect, then, as Abramovitzation of effective demande discernible long swings.

of the appropriate PART II / FERTILITY ANALYSESto variables and relation-en customary in the past.

—

r

CHAPTER 4 / THE AMERICAN BABYBOOM IN HISTORICAL PERSPECTIVE

This chapter focuses on the recent baby boom, particularly on theextent to which this represents an abrupt break with past experience.The next chapter extends the analysis to the current fertility decline.

We first take a fresh look at the historical record in the light ofthe Kuznets-cycle conception of economic change, taking care todiscuss separately the experience of three population groups withsignificantly different patterns—foreign-born, native-born urban, andnative-born rural. After a brief retracing of several earlier findings,the analysis quickly moves onto new ground, exploring some possiblereasons for the pattern observed.

The analysis is confined to the white population because of thegreater reliability of the data for this group and its predominantinfluence in determining the pattern for the total.

KUZNETS CYCLES IN U.S. POPULATIONGROWTH AND FERTILITY

The Rate of Total IncreaseWe start with the rate of population growth. Since we are

interested in focusing on major movements, we employ five-yearaverages of the basic data,' a choice governed partly by preference—to eliminate or at least reduce the shorter-term changes associatedwith the ordinary business cycle—and partly by necessity—because ofthe initial mold in which some of the basic data are cast, particularlythose relating to fertility.

NOTE: An earlier version of this chapter was published in December 1981 in theAmerican Economic Review, and was reprinted by the National Bureau as itsOccasional Paper 79.

'For the rate of total increase, the average is implicit. The rate, which Isactually calculated from observations on the population stock separated by fiveyears, yields a time pattern equivalent to that of a geometric average of the annualrates of change within the successive quinquennia.

.

78 / II: FERTILITY ANALYSES

Figure 22 shows the average rate of increase of the U.S. whitepopulation in successive quinquennia from 1870 to 1955. The familiardownward drift through the 1930's and the recent increase areimmediately apparent. Less familiar, but equally obtrusive, are sig-nificant fluctuations in the rate of change. The duration of thefluctuations has run from ten to thirty-five years and their averagemagnitude has amounted to about one-quarter of the mean rate ofchange over the period as a whole. In Chapter 2, the components ofpopulation growth responsible for these fluctuations were identified(Figure 2). It was noted that, while immigration was typically theprincipal factor in the past, the recent population upsurge has beendue to fertility.

Since 1870, then (and indeed even before [103, p. 36; 107]), the

FIGURE 22AVERAGE GROWTH RATE OF TOTAL WHITE POPULATION, 1870—1959

SoURcE: Table C-i.

historical record has consthe rate of population grupsurge in the rate of pbirth rate rather than inrecent increase bears onlythat, given the new immin the rate of growth waris correct or whether the rto its forebears is a questichapter.

The Birth Rate of tJLet us turn to the con

the center of our intereit possible to reconstrucithe white population ofestimates have been avEkeeping with our interest

The upper panelof

the movemeby a five-year average—.through the secular troualternated with intervalsThese are the longsomewhat different set centire eighty-year perioc

2 Economists are perhapson population change. Wheino annual data on the crudeperspective to the years befoilion by Melvin Zelnik, carrliUniversity, under the supersrived by applying appropriatof-age distributions adjustedages, primarily those endingthe patterns traced by

the same; for earlielreliable because of the lesseimates. The dissertation hasthe figures in the thesis.

1875- 85- 95- 05- 15- 25- 35- 45-80 90 00 10 20 30 40 50 59

THE AMERICAN BABY BOOM / 79

increase of the U.S. white1870 to 1955. The familiarthe recent increase are

equally obtrusive, are sig-ge. The duration of thee years and their averagearter of the mean rate ofpter 2, the components ofuctuations were identifiedigration was typically the

upsurge has been

re [103, p. 86; 107]), the

OPI.TLATION, 1870—1959

historical record has consistently been marked by major swings inthe rate of population growth. But since the source of the recentupsurge in the rate of population growth has been a rise in thebirth rate rather than in immigration, one might maintain that thisrecent increase bears only a surface resemblance to prior swings andthat, given the new immigration restrictions of the 1920's, recoveryin the rate of growth was hardly to be expected. Whether this viewis correct or whether the recent movement does bear a logical relationto its forebears is a question to which I return toward the end of thechapter.

The Birth Rate of the Total White PopulationLet us turn to the component of population change that constitutes

the center of our interest, the birth rate. Recent work has madeit possible to reconstruct a full century of fertility experience forthe white population of the United States.2 The annual birth rateestimates have been averaged here for successive quinquennia, inkeeping with our interest in discerning Kuznets cycles.

The upper panel of Figure 23 brings out clearly the long-termdecline in the level of the birth rate and its recent recovery. It alsoshows that the movement of the birth rate—even when smoothedby a five-year average—has been far from regular. For the periodthrough the secular trough of the 1930's, intervals of rapid declinealternated with intervals of slower decline or even absolute increase.These are the long swings in fertility which Kuznets found in asomewhat different set of figures. They are apparent throughout theentire eighty-year period of fertility decline covered here.

2 Economists are perhaps not generally aware of the scarcity of historical dataon population change. When Kuznets made his study in 1958 [103], there wereno annual data on the crude birth rate before 1909. The new series, extending ourperspective to the years before the Civil War, is the product of a doctoral disserta-lion by Melvin Zelnik, carried on at the Office of Population Research, PrincetonUniversity, under the supervision of Ansley Coale [2201. The estimates were de-rived by applying appropriate mortality rates to the decennial census single-year-of-age distributions adjusted for "age heaping" (excessive reporting of certainages, primarily those ending in 0 and 5). As the upper panel of Figure 23 shows,the patterns traced by these and the official estimates in the overlap period arevirtually the same; for earlier dates, however, the Zelnik figures are somewhat lessreliable because of the lesser accuracy or availability of data needed for the esti-mates. The dissertation has now been published [321, but my analysis above usesthe figures in the thesis.

The lower panel ofcentage rate of changethe data plotted in theper quinquennium8 per cent. If this rate hiperiod, the individual obline shown in the figureabout the line makes cliwere of substantialdeviations from the meadecline itself. The duratiof this century was fifteto peak or trough to tnof much longer duration

But of what interesitative history forsuggest that it leads to 1ord, which has significaence. Typically, the histcthe long-term secular detion that this declinebeen repeatedly charactand substantialleads naturally to theabrupt break—a reversalof historical change enmight be conceived as Iaround the trend—agreater amplitude andview implies less of araises the possibility of,

To avoid confusion, it Snot the rate of change therechange shown in Figure 22not necessarily implycohorts (see the concluding'

For examples of this see153, Ch. 13].

FIGURE 23LEVEL AND RATE OF CHANGE OF CRUDE BIRTH RATE OF TOTAL WHITE

POPULATION 1855—1959Crude birth rate(per thousand population)

Ratio scale

40 -

20 -

I I I I

55-59

1855- 65- 75- 85- 95- 05- IS- 25— 35- 45-59 69 79 89 99 09 19 29 39 49

1855-59 65-69 75-79 15-19 25-29 35-39to to to to to to to to to to

60-64 70-74 80-84 90-94 00-04 10-14 20-24 30-34 40-44 50-54

SouRcE: Table C-2.

R RATE OF TOTAL WHITE)59

-. - -- -'

25— 35- 45-29 39 49

55-59


The lower panel of Figure 23 presents the quinquennial per-centage rate of change of the birth rate, computed directly fromthe data plotted in the upper The average rate of declineper quinquennium through the secular trough in 1935—39 was about6 per cent. If this rate had prevailed uniformly throughout the entireperiod, the individual observations would have formed the horizontalline shown in the figure. The movement in the actual observationsabout the line makes clear that the variations in the rate of changewere of substantial magnitude; in fact, the average value of thedeviations from the mean amounts to six-tenths of the mean rate ofdecline itself. The duration of the two swings through the first decadeof this century was fifteen to twenty years, whether measured peakto peak or trough to trough. The movements since then have beenof much longer duration, on the order of thirty-five to forty years.

But of what interest, it may be asked, is this exercise in quanti-tative history for analysis of the baby boom? In reply, one mightsuggest that it leads to revising one's conception of the historical rec-ord, which has significant implications for interpreting recent experi-ence. Typically, the historical movement which has been emphasized isthe long-term secular To this I would now add the observa-tion that this decline has been far from regular; that, in fact, it hasbeen repeatedly characterized by fluctuations of noticeable amplitudeand substantial duration. The customary interpretation of the pastleads naturally to the view that recent experience constitutes anabrupt break—a reversal in primary trend. In contrast, the conceptionof historical change employed here suggests that recent experiencemight be conceived as the latest in a succession of major movementsaround the trend—a Kuznets cycle which, for some reason, is of muchgreater amplitude and duration than its predecessors. Clearly thisview implies less of a break with historical experience and at leastraises the possibility of more easily reconciling the present with the

To avoid confusion, it should be noted that (1) it is the birth rate itself andnot the rate of change therein that is the component of the rate of total populationchange shown in Figure 22, and (2) swings in annual birth or fertility rates donot necessarily imply swings in the completed fertility of successive populationcohorts (see the concluding section of Chapter 5).

For examples of this see [86, 185, 206] and more recently [69; 82, Chs. 2, 11;153, Ch. 13].

LL Iii25-29 35-39 45.49

to30-34 40-44 50-54


past—a sine qua non of any attempted explanation of the baby boom.Moreover, it suggests a new research strategy with regard to thebaby boom, namely, that one focus on explaining Kuznets cycles, pastand present, in an effort to determine whether the underlying causesof these movements may have operated with exceptional force inrecent decades. It is in terms of this conception that the subsequentanalysis is organized.

Before proceeding to this analysis, there is one more feature ofFigure 23 that deserves attention. This is the precipitous decline inthe birth rate during the 1920's. A trend line fitted to the pre-1920data in the upper panel and extended through the next two decadeswould lie not only above the observations for the 1930's, but abovethat for 1925—29 as well. From the lower panel, one finds that therate of decline between the first and second halves of the 1920's wasthe second highest in the one hundred-year record, falling onlyslightly below that in the next overlapping decade. This drastic declineduring a period of high prosperity has been cited by demographersas grounds for discounting efforts to explain the baby boom on thebasis of economic factors. For example: ". . . the interpretation ofthe baby boom as the natural consequence of prolonged prosperityis hardly more tenable than the earlier interpretation of the reversalin the 1930's as momentary. The next earlier period of notable prosper-ity in the United States—the 1920's—was a period of sharply fallingfertility. In fact, as Dudley Kirk points out, the depressed 1930'sproduced more births by far than one would expect on the basis ofan extrapolation of the trend of the prosperous 1920's."

Ansley J. Coale, Introduction [170, pp. 5—6]. The reference is to Dudley Kirk,"The Influence of Business Cycles on Marriage and Birth Rates" [170, pp. 241—260]. The method followed by Kirk in his analysis is to correlate "trend deviationsof economic measures (as independent variables) to measures of nuptiality andnatality (as dependent variables)" [170, p. 242], using fertility data for the totalpopulation for the period 1920—58. While the results are relevant to analysis offertility variations within the ordinary business cycle, in my view they cannot beused to draw inferences about the baby boom. The "trend" lines fitted for theperiod 1920—58 largely reproduce the Kuznets cycle which constitutes the babyboom. By concentrating on explaining deviations from "trend," Kirk in effecteliminates from his analysis the baby boom itself. Moreover, even with regard tobusiness cycle analysis, it would be of interest to distinguish components of thetotal population whose fertility was subject to substantially different influences, asis done below for Kuznets cycles.

TABLE 1. PERCENTAGE20-44, BY NATIVITY,BY RURAL-URBAN RESII

Total whiteForeign-born whiteNative white

UrbanRural

SOURCE: Table C-5 and cera Based on 1950 census defi

Clearly, an attempt tdevote special attention

The Fertility of theThe fertility of the I

of a number ofto common, influences. \boom if we considerforeign-born white popiand rural components. Tfemales of reproductivevarious dates. In the pifor the foreign-born and

For the dependentsuse the fertility ratio, Uthe number of womenthe available data.° As t]typically exceeds the ciiof twenty to twenty-five:

Total WhiteCrude birth rate (annulFertility ratio (next cens

6 A good discussion of thfertility ratio is given in [&

nation of the baby boom.Ltegy with regard to theining Kuznets cycles, pasther the underlying causesvith exceptional force inption that the subsequent

re is one more feature ofthe precipitous decline inme fitted to the pre-1920igh the next two decadesfor the 1930's, but abovepanel, one finds that thehalves of the 1920's was

ear record, falling only:cade. This drastic declinea cited by demographersa the baby boom on the

• . the interpretation ofof prolonged prosperity

pretation of the reversalieriod of notable prosper-period of sharply fallingut, the depressed 1980'sd expect on the basis ofus 1920's."


TABLE 1. PERCENTAGE DISTRIBUTION OF WHITE FEMALES,20-44, BY NATIVITY, AND OF NATIVE WHITE FEMALES, 20-44,BY RURAL-URBAN RESIDENCE, 1890-1950

1890 1910 1930 1950

Total white 100.0 100.0 100.0 100.0Foreign-born white 20.9 19.9 14.7 4.6Native white 79.1 80.1 85.3 95.4

Urban 30.2 39.6 51.5 64.7Rural 48.8 40.5 33.8 30.7

SOURCE: Tabk C-5 and census reports.a Based on 1950 census definition of "urban."

Clearly, an attempt to reconcile present with past experience mustdevote special attention to the record for the 1920's.

The Fertility of the Native and Foreign-Born White PopulationsThe fertility of the total white population is a composite of that

of a number of subgroups, each subject in part to distinctive, in partto common, influences. We can gain further perspective on the babyboom if we consider separately the experience of the native andforeign-born white populations, and, within the former, the urbanand rural components. Table 1 indicates the proportion of total whitefemales of reproductive age accounted for by each of these groups atvarious dates. In the present section, we consider fertility patternsfor the foreign-born and total native white populations.

For the dependent variable, instead of the crude birth rate I nowuse the fertility ratio, the number of children under 5 years old tothe number of women 20 to 44 years old, a choice necessitated bythe available As the following figures suggest, the fertility ratiotypically exceeds the crude birth rate by a factor in the neighborhoodof twenty to twenty-five:

Total White Population 1885—89 1905—9 1925—29

Crude birth rate (annual average) 35.3 29.4 22.4Fertility ratio (next census date) 744 632 505

° A good discussion of the conceptual and statistical problems relating to thefertility ratio is given in [82, p. 13 and Appendix A].

reference is to Dudley Kirk,Birth Rates" (170, pp. 241—to correlate "trend deviations

measures of nuptiality andng fertility data for the totals are relevant to analysis ofin my view they cannot be"trend" lines fitted for thewhich constitutes the baby

mm "trend," Kirk in effectoreover, even with regard toIstinguish components of the

different influences, as

I


Analytically, this reflects the fact that the fertility ratio is computedfrom (a) a denominator about one-fifth as large as that for thecrude birth rate (females aged 20—44 instead of the total popula-tion), and (b) a numerator four to five times as large. (Implicitly,birth experience over a five-year period is totaled rather than averaged,and is multiplied by a survival rate on the order of .85 to .95 to ex-clude those dying before the end of the period.) Thus the time pat-terns traced by the two measures may differ somewhat because ofvariations in the ratio of women aged 20—44 to the total populationand in the mortality of children under 5 years, particularly in infantmortality.7

Figure 24 presents fertility ratios for the total white populationby nativity from 1875—79 to 1925—29, and, supplemented by generalfertility rates, for the native and total white populations to 1954—58.8 The observations on fertility ratios are at census and mid-censusdates, but since they reflect fertility behavior over the preceding fiveyears, we have dated them according to the quinquennia to which theyrefer. The lower panel shows the percentage rate of change perquinquennium in each series, computed in the same fashion as for thepreceding figure.

Several points deserve mention. First, Kuznets cycles are evidentin the series for both the native- and foreign-born groups. Through1925—29, the timing of the swings appears to be usually the same,but the amplitude is substantially greater for the foreign-born white.There is some suggestion of increasing amplitude, particularly for thenative white, and in the most recent period the magnitude of theswing for this group is strikingly greater than previous ones. Arithmetic

LEVEL AND RATE OF CHANGEBAL FERTSLITY StATE, 192

Fertility ratio1.000

900 -

600 —

500 -

400 -

300 I I I

1865- 75- 85-69 79 89

V. change ptr Quinquenniumintertility ratio

1 For the total white population, the only one for which comparison is possible,the directions of change in the rate of change of the crude birth rate and of thefertility ratio are identical from 1885—89 on, the principal period of the analysis,with one exception. This exception is primarily due to an understatement of thefertility ratio for 1910—14, because no adjustment was made for the exceptionaleffect of the influenza epidemic of 1918.

8 The fertility ratio estimates, prepared in connection with the present study,are based in large part on a valuable unpublished memorandum prepared byEverett S. Lee providing age and parentage detail underlying the quinquennialestimates of native white population published by Kuznets [103]. Because ofomissions or defects in the recent reporting of parentage and nativity, it was notpossible to continue these estimates beyond 1925—29. However, to provide someidea of the pattern after 1925—29, use has been made of the official estimates ofthe closely comparable general fertility rate (live births per 1,000 females aged15—44) for the total and native white populations.

+10 —

a

0-

—30 I

1870-74 80-84 90-to to I

75-79 85-89 95

Souisca: Table C-3.

analysis of the swingscaused in important minative and foreign-borshifts in the relative imi

r


fertility ratio is computedas large as that for thehead of the total popula-mes as large. (Implicitly,:aled rather than averaged,order of .85 to .95 to ex-nod.) Thus the time pat-Fer somewhat because of4 to the total population 120

ars, particularly in infant 110

100

ie total white population 9°

supplemented by general 80

ite populations to 1954— 70at census and mid-censusr over the preceding fiveuinquennia to which theyage rate of change pere same fashion as for the

uznets cycles are evidentn-born groups. Throughto be usually the same,

the foreign-born white.tude, particularly for the

the magnitude of the)revious ones. Arithmetic

vhich comparison is possible,crude birth rate and of the

icipal period of the analysis,to an understatement of theas made for the exceptional

ion with the present study,memorandum prepared by

tnderlying the quinquennialKuznets [103]. Because ofsge and nativity, it was not

However, to provide someof the official estimates of

ths per 1,000 females aged

FIGURE 24LEVEL AND RATE OF CHANGE OF FERTILITY 1865—1929, AND OF GEN-

ERAL FERTILITY RATE, 1920—58; TOTAL WRITE POPULATION, BY NATIVITYFertihty rotLo General fertility rote1,000

Ratio scales 0101 whtte(scale—) —

I I I

1865- 75- 85- 95- 05- 15-69 79 89 99 09 19

I I I _I I I I I I

25— 35- 45- 54-29 39 49 58

1870-74 80-84 90-9400-04 10-14 20-24 30-34 40-44 50-54to to to to to to to tO to

75-79 85-89 95-99 05-09 15-19 25-29 35-39 45-49 54-58

SouRcE: Table C-3.

analysis of the swings in the total white group shows that they arecaused in important measure by the fertility movements of both thenative and foreign-born components, and that the contribution ofshifts in the relative importance of the two groups has been negligible.

tI


The native-born white group, despite the smaller amplitude of itsswings, typically accounted for the dominant part of the movementin the total because of its much greater share (Table C-12).

Some light is also cast on the precipitous rate of decline in totalwhite fertility in the 1920's. For both the foreign- and native-bornpopulations there is a substantial drop in the fertility ratio between thefirst and second halves of the decade. However, the decline for theforeign-born is more than double that for the native—29 against 12per cent. Hence, a significant part of the decline in total white fertilityin the 1920's—to be precise, about one-third (Table C-12)—was owingto the drastic reduction in the fertility of the foreign-born white pop-ulation.° Indeed, for this group, if one adds the movement between thetwo preceding quinquennia, the drop in fertility was nothing short ofspectacular. Between 1915—19 and 1925—29 the foreign-born whitefertility ratio dropped by about four-tenths, more than double thedecline during the preceding forty years.

The Fertility of the Urban and Rural Native White PopulationsOur data now become even more limited, relating only to the

latter half of each decade from 1885—89 on. Estimates published by theNational Resources Committee [204] for 1905—9 through 1925—29 havebeen carried back two additional decades. A constant 5 per cent ad-justment by the NRC for underenumeration of children under fiveyears has been accepted here, in part because no basis for a differen-tial rural-urban adjustment was readily available, and in part becausethe analysis rests primarily on the figures for the more reliable censusesfrom 1900 on. Our immediate interest is in the pattern through 1925—29, and estimates for the native white population by rural-urban resi-dence are only available to this point. To fill out the picture since then,however, we have added overlap figures for the total white populationfor 1925—29 on, an approximation which seems reasonable in view ofthe much diminished importance of the foreign-born in recent years.

As is clear from the curve for the total native white group inFigure 25, compared with that in Figure 24, the timing of the Kuznetscycles before 1925—29 is such that omission of the observations for

"The decrease in fertility of foreign-born white women was perhaps the out-standing feature of the decline in the birth rate during the twenties" [205, p. 127].

LEVEL AND RATE OF CRANDENCE: NATIVE WRITE POi't

Fertihty ratio1,000

Ratio scale900-Rurøl native white800-

700 —

500 —

400- '-..

SOC885-89 95-99 05-09

°6 change oer decodein fertility ratio+50

+40 —

+30 —

+20 —

— Ruroirtotivewhite

/0_I ./-

—10 — Totatnt

—20 —

—30

_______________

885-89 95-99to to

95-99 05-09

SouacE: Table C-6.

Ir

FIGURE 25

smaller amplitude of itsant part of the movementIre (Table C-12).isis rate of decline in totalforeign- and native-bornfertility ratio between the

ivever, the decline for thethe native—29 against 12

in total white fertility(Table C-12)—was owinge foreign-born white pop-he movement between thetility was nothing short of9 the foreign-born white

more than double the

Vative White Populationsited, relating only to thestimates published by the

5—9 through 1925—29 haveA constant 5 per cent ad-in of children under fivese no basis for a differen-lable, and in part becausethe more reliable censuses:he pattern through 1925—ation by rural-urban resi-out the picture since then,the total white populationms reasonable in view ofign-born in recent years.

al native white group inthe timing of the Kuznets

of the observations forwomen was perhaps the out-ig the twenties" [205, p. 127].

L

LEVEL' AND RATE OF CHANCE OF FERTILITY RATIO BY RURAL-URBAN RESI-DENCE: NATIVE WHITE POPULATION, 1885—1929; TOTAL WHITE POPULATION,

1925—58Fertility ratio1,00c - — -

900

800

700

600

Ratio scale

• Rural native white

- /Rural total white

%

500 —

400

w8ite

total white

885-89 95-99 05-09 1519 35-39 45-49 54-58

to to95-99 05-09

SoURcE: Table C-6.


the first half of each decade tends to conceal the long swings. Never-theless, some significant points stand out. As the upper panel shows,the decline from 1885—89 to 1925—29 in fertility of the total nativewhite population was significantly greater than that for either of thecomponents. This was caused by the depressing influence on totalnative white fertility of the continuous redistribution of populationfrom high-fertility rural to low-fertility urban areas. Quantitativelythis rural-urban shift accounted for about one-half of the total declineover the forty-year period (Table C-iS). The depressing effect onfertility was about the same in each successive decade.

A second point of interest is the greater decline in rural thanurban fertility through 1925—29. The rural decline is about half againas great as the urban—18 as opposed to 12 per cent. Indeed, if oneconsiders the estimates for urban fertility from only 1895-99 to 1925—29, there is little evidence at all of a declining trend. The over-allreduction in these three decades is only 4 per cent, and the impressioncreated by the curve is one of general stability.

This observation of substantial stability for a group accounting inthis period for a third to a half of white females of reproductive ageruns so counter to the common impression of a general and persistentsecular decline that it deserves further consideration. This is particu-larly true since this group has tended to assume an increasinglydominant role in determining the pattern for the total white populationand thus is of central significance for consideration 0f recent andprospective experience of the white population as a whole.'° Couldthe finding be a statistical artifact resulting from deficiencies in ourestimating procedure? The possibility cannot be discounted—I haveattempted to make a reasonable estimate for 1895—99, but with moretime and larger resources it undoubtedly could be improved. However,even if we take only the more firmly based NRC estimates for 1905—9through 1925-29—at the expense unfortunately of reducing our spanof observation to two decades—there is still little evidence of a signifi-cant decline. In presenting these data the NRC does not call into

10 Readers may be reminded in this connection of the finding in Dorothy S.Thomas' pioneering study of Sweden [1621 that during the nineteenth centuryshort-term fluctuations in fertility of the total population were initially dominatedby fluctuations in agriculture, but subsequently by those in industry.

question the figures forare accorded hardly anregional fertility patternNRC does note that "thleveling off of birth raterate pattern" [205, p. ii

Some additional hisstability is perhaps wort]a study of the fertilityEngland, in which he co:and 1915 no definite treip. 34]. For the period Istudy), he found anarea in the forefront oftion in which native whlong period stretchingappearance of a similardate would clearly beexperience.

One final point shoiof total native white femore to a decrease in n1925—29, the reduction ithat for urban fertilitying of this perioddecline.

SummaryWhile the fertility

tially from the latter pithere was significantamong componentas to eliminate or subscycle, marked fiuctuat

11 A reexamination by Rsetts birth registration dataent evaluation from Spengi


question the figures for urban native white population, though theyare accorded hardly any attention [205, p. 1271. With regard toregional fertility patterns of the total white population, however, theNRC does note that "these data show clearly a tendency toward theleveling off of birth rates in areas long influenced by the lower birthrate pattern" [205, p. 128].

Some additional historical evidence consistent with the finding ofstability is perhaps worth citing. In 1930, Joseph J. Spengler publisheda study of the fertility of native- and foreign-born women in NewEngland, in which he concluded that "during the period between 1860and 1915 no definite trend appeared in the native fertility rates" [149,p. 84]. For the period from 1915 through 1925 (the last year of thestudy), he found an upward tendency in fertility. Here, then, is anarea in the forefront of the process of urbanization and industrializa-tion in which native white fertility did not significantly decline over along period stretching well back into the nineteenth century.11 Theappearance of a similar pattern for the nation as a whole at a laterdate would clearly be consistent with this earlier New Englandexperience.

One final point should be noted regarding Figure 25. The declineof total native white fertility in the 1920's is now seen to be owingmore to a decrease in rural than urban fertility. Between 1915—19 and

• 1925—29, the reduction in rural fertility was close to 10 per cent, whilethat for urban fertility was under 6 per cent. Thus further understand-

• ing of this period calls particularly for an explanation of the ruraldecline.

SummaryWhile the fertility of the total white population declined substan-

tially from the latter part of the nineteenth century to the mid-1930's,there was significant variation in the rate of change over time andamong component population groups. Even after averaging data soas to eliminate or substantially reduce variability due to the businesscycle, marked fluctuations—Kuznets cycles of fifteen or more years

"A reexamination by Robert Gutman [83] of the reliability of the Massachu-setts birth registration data used by Spengler, while arriving at a somewhat differ.ent evaluation from Spengler, does not upset this finding.

L

p.,

1 the long swings. Never-s the upper panel shows,tiity of the total nativeian that for either of the

influence on totalistribution of populationan areas. Quantitatively

of the total declinedepressing effect on

sive decade.er decline in rural thandine is about half againper cent. Indeed, if oneiii only 1895—99 to 1925—rung trend. The over-allcent, and the impression

r a group accounting inales of reproductive agea general and persistent•eration. This is particu-assume an increasinglyre total white population

on a whole.'° CouldTom deficiencies in ourbe discounted—I have

1895—99, but with morebe improved. However,

estimates for 1905—9y of reducing our span:le evidence of a signifi-RC does not call intothe finding in Dorothy S.

ing the nineteenth century)n were initially dominatedse in industry.


duration—stand out in the patterns for the total, native, and foreign-born white populations. Moreover, in the first three decades of this cen-tury the over-all decline in total white fertility was owing almost ex-clusively to declines for the foreign-born white and rural native whitepopulations and to the shift from rural to urban areas; the fertilityof the urban native white population, the group of central importancein understanding recent and prospective movements in the aggregate,remained virtually unchanged. Considerations such as these raise thequestion whether the baby boom, rather than an abrupt reversal ina long-term downtrend, might not be at least in part a Kuznets cycleof much larger magnitude than heretofore. To answer this, it isnecessary to look into possible reasons for these movements.

REASONS FOR KUZNETS CYCLES IN FERTILITYOF DIFFERENT POPULATION GROUPSBriefly stated, the analytical viewpoint underlying the subsequentdiscussion is this: variations in the fertility of a given population groupare caused primarily by changes in two classes of factors—economiccondition and demographic composition. The "group" for which thesefactors should be studied comprises those in the family-building ages.Broadly, this embraces those aged 15—44 years, but for some purposesparticular attention should be paid to the younger members, thoseaged, say 20—29, where so many decisions regarding marriage andchildbearing are concentrated. "Economic condition" refers to theemployment and income experience of the group. Ideally, "income"here would embrace all sources, including even interpersonal transfersfrom other age groups, though in the following discussion attentionis concentrated on the chief source, labor income. "Demographiccomposition" refers to the distribution of the group, according tocharacteristics such as age, sex, nationality, and parentage. A changein demographic composition may itself stem basically from economicforces, for example, a change in age composition of the foreign-borndue to a rise in immigration, but it is nevertheless useful to distinguishthe different channels through which these forces operate. Botheconomic condition and demographic composition may affect theover-all fertility of a population group by influencing either marriagebehavior, marital fertility, or both. No consistent effort is made here

to distinguish the role ofchange, though it would

The analysis belowtional factors, while thoston economic condition. I

where possible, the inflifertility in so far as theychange, and of changesfertility, especially those.this discussion, however,.,but have singled out tho'initial investigation to t]shown by each group.

Foreign-Born White'As populations go,

because the source of itiOne result of this is aage distribution of afall progressively with eishows a concentration innumbers at the extremes,[164, p. 144}. Moreover,fed in at relatively advthere is a significant di

I

12 This brief statement ofthe determinants studied hestions in the competitive situl,associated, e.g., with the mtthe net income which childa56—58] and Gary S. Becker'credit resources; and shiftsother contributions by econolalso my recent paper cited is

13 A cursory look at the awhite rural and urban populdecade variability than theThompson and Wheipton [Iand 27]. While there are sos[19, 153], it is unfortunatestudy to the present in its

14 Children born to foreigsfled as native-born.

-

TEE AMERICAN BABY BOOM / 91

to distinguish the role of these two components in over-all fertilitychange, though it would be of interest in a fuller treatment.12

The analysis below for the foreign-born takes up only composi-tional factors, while those for the two native-born groups concentrateon economic condition. It would have been of interest to examine,where possible, the influence of economic factors on foreign-bornfertility in so far as they exert effects other than through compositionalchange, and of changes in demographic composition on native-bornfertility, especially those associated with rural-urban migration.18 Inthis discussion, however, I have not attempted an exhaustive analysis,but have singled out those factors which seemed on the basis of myinitial investigation to throw significant light on the Kuznets cyclesshown by each group.

Foreign-Born White FertilityAs populations go, the foreign-born is an unusual one—primarily

because the source of its growth is immigration rather thanOne result of this is a very atypical age, distribution. Unlike the usualage distribution of a growing population, where the numbers tend tofall progressively with each older age group, that of the foreign bornshows a concentration in the middle age groups with relatively smallnumbers at the extremes, at least as long as immigration remains high[164, p. 144]. Moreover, not only are the additions to this populationfed in at relatively advanced ages—the "prime" working ages—butthere is a significant disproportion between the sexes, with males

—

:otal, native, and foreign-three decades of this cen-ity was owing almost ex-ite and rural native whiteurban areas; the fertilityup of central importance'ements in the aggregate,iS such as these raise thean an abrupt reversal int in part a Kuznets cycle

To answer this, it isiese movements.

IN FERTILITYLOtJPS

clerlying the subsequenta given population group

of factors—economic"group" for which these

the family-building ages.but for some purposes

younger members, thoseregarding marriage andcondition" refers to thegroup. Ideally, "income"en interpersonal transfersiring discussion attention

income. "Demographicthe group according tond parentage. A changebasically from economicition of the foreign-bornless useful to distinguishe forces operate. Both

may affect theuencing either marriagetent effort is made here

12 This brief statement of analytical viewpoint is intended merely to highlightthe determinants studied here. Among other possibly important factors are varia-tions in the competitive situation of children in the consumers' scale of preferenceassociated, e.g., with the introduction of new consumer durables, or a change inthe net income which children add to the family (see Joseph S. Davis [45, pp.56—58] and Gary S. Becker [170, pp. 209—2.31]); changes in the availability ofcredit resources; and shifts in techniques and knowledge of birth control. Forother contributions by economists, see [2, 7, 84, 112, 132, 134, 214 and 145a]. Seealso my recent paper cited in Chapter 5, footnote 5.

13 A cursory look at the available data on compositional aspects of the nativewhite rural and urban populations suggests that they exhibit much less decade-to-decade variability than the foreign-born white. See the 1890—1930 figures inThompson and Whelpton [164, Tables 41 and 56, and Appendix Tables 17, 23,and 27]. While there are some excellent recent general studies on U.S. population[19, 153], it is unfortunate that there is nothing that continues this remarkablestudy to the present in its full analytical depth.

14 Children born to foreign-born women after immigration are, of course, classi-fied as native-born.


noticeably predominating. Finally, given wide swings in immigration,such as have occurred in this country, the relative size even ofadjoining age-sex groups can fluctuate widely in as short a period asa decade.

These considerations explain my starting with demographic com-position in seeking clues to the variations in the rate of change offoreign-born fertility. My immediate point of departure in studyingthese movements, particularly the very steep decline in the 1920's wasthe observation that the proportion of young foreign-born women whowere married dropped sharply from 192() to 1930, as is shown by thefollowing figures:

Age at Percentage MarriedSpecified Date

Why, one may ask, should such an abrupt decline occur? The chancethat a foreign-born white woman aged 20—24 by 1920 was married wasalmost two in three, but if she reached this age group only one decadelater, the likelihood had declined to less than one in two.

An obvious hypothesis, stemming from the observation that theproportion of young foreign-born men married remained almostconstant over the decade, is that the demand for women to marrydropped off because of a decline in the relative number of males inthe market [101, 166]. In testing this, however, one must recognizethat the relevant ratio is not that of males to females in a given agegroup, the standard sex ratio, since, as is well known, men typicallymarry at a later age than women. For example, in the period 1890—1930, at least 45 per cent of foreign-born white women were marriedby the time they were 20—24, but for foreign-born white men thisproportion was not attained until ages 25—29 had been reached [164,p. 395]. In attempting to explain the marriage proportion for foreign-born white women aged 20—29, therefore, the ratio of foreign-bornwhite males aged 25—34 to females aged 20—29 was

15 The analysis implies of course that native-born men did not constitute aparticularly important source of demand for foreign-born women. This assumptionseems consistent with the facts; in 1920 the proportion of foreign-born motherswhose husbands were native-born was less than one in six [188, p. 28Z].

20—2425—29

192061.681.6

193047.575.9

RATIO OF MAJ20—29, AND

EICN-BOH1

Per cent200

ISO —

100 —

90 —

80 —

70 —

60

501890

Ta]

The relevant seriecensus datestraced by the two cuthe standard sex ratioApparently, the maritaldid depend very conorthodox sex ratio, whiand current immigratitpopulation.1°

In Figure 27,Here, at five-year inte

16 An interesting by-pnratio during the 1920's,foreign-born white femalethe labor-force participathA-4].


e swings in immigration,Le relative size even of

in as short a period as

with demographic corn-i the rate of change ofI departure in studyingdecline in the 1920's wasoreign-born women whoL930, as is shown by the

age Married193047.575.9

dine occur? The chance1920 was married was

s group only one decaden one in two.he observation that therried remained almostd for women to marryive number of males in'er, one must recognizefemales in a given age

II known, men typically1e, in the period 1890—

te women were marriedrn-born white men thishad been reached [164,proportion for foreign-

e ratio of foreign-born9 was

men did not constitute arn women. This assumptionon of foreign-born motherss six t188, p. 232.).

-

FIGURE 26RATIO OF MALES AGED 25—34 TO FEMALES AGED20—29, AND PERCENTAGE OF LATTER MARRIED: FOR-

EIGN-BORN WHITE POPULATION, 1890—1930Per cent

150

100 —

90 —

80 —

Females 20 29

Ratio scale

1890 1900 1910 1920 1930

SouRcE: Table C-7.

The relevant series are plotted in Figure 26 for the decennialcensus dates 1890—1930. The close similarity between the patternstraced by the two curves—a similarity which would not appear ifthe standard sex ratio for those aged 20—29 were used—is impressive.Apparently, the marital experience of young foreign-born white femalesdid depend very considerably on the gyrations of our rather un-orthodox sex ratio, which in turn arose from the impact of both earlierand current immigration on the age-sex structure of the foreign-bornpopulation.18

In Figure 27, this line of reasoning is pushed a step further.Here, at five-year intervals, the series for foreign-born white fertility

16 An interesting by-product of the sharp decline in the marriage-relevant sexratio during the 1920's, and the corresponding reduction in the proportion offoreign-born white females aged 20—24 who were married, was an abrupt rise inthe labor-force participation of this group from 37.6 to 50.1 per cent [118, TableA-41.


FICTJRE 27

LEVEL AND RATE OF CHANCE OF FERTILITY RATIO; AND OF RATIO OF MALESAGED 25—34 TO FEMALES 20—29, AND OF FEMALES AGED 20—34 TO FEMALES

20—44: FOREIGN-BORN WHITE POPULATION, 1875—1930

and our marriage-releva]the latter being used imarriage proportion atwhile the movements Inot perfectly consistentshow two trough-to-trotclose, if not identical.sponsible for Kuznetsfertility was the changiaged 20—29 and the coition.

The broken line iifeature of the foreign-ithe fertility swings, naprime reproductive

at

thesome

This brief discussiforeign-born white ferdefinitive, and enoughfor the view that shifborn associated with tiin part responsible fosuggestion, however,One—of particular mtof the period conside:fertility was somewhabasis of the twosuggested in several scompositionalproportion of foreign-Icame from the high-feThere is substantial etypically had

80-84 90-94 00-04 10-14

SOURCE: Table C-8.

TTHE AMERICAN BABY BOOM / 95

and our marriage-relevant sex ratio (the two solid lines) are compared,the latter being used in the absence of direct observations on themarriage proportion at mid-census dates. As the lower panel shows,while the movements in the rates of change of the two series arenot perfectly consistent, there is a noticeable similarity. Both seriesshow two trough-to-trough swings with the dates of peaks and troughsclose, if not identical. This suggests that at least one element re-sponsible for Kuznets cycles in the rate of change of foreign-bornfertility was the changing proportion of males aged 25—34 to femalesaged 20—29 and the consequent effect of this on the marriage propor-tion.

The broken line in the figure brings out a second demographicfeature of the foreign-born population that may have contributed tothe fertility swings, namely, the proportion of women aged 20-44 inprime reproductive ages, conceived here as encompassing ages 20—34.Here too there is a suggestion of two trough-to-trough swings withreasonably consistent timing, though the amplitude of the movementsis somewhat smaller for this series. However, in the beginning of theperiod (for which the estimates are probably less reliable), the timingrelationships are somewhat off.

This brief discussion of Kuznets cycles in the rate of change offoreign-born white fertility is designed to be exploratory rather thandefinitive, and enough has perhaps been said to provide some supportfor the view that shifts in demographic composition of the foreignborn associated with the changing impact of immigration were at leastin part responsible for these movements. Even if one accepts thissuggestion, however, there remain some troublesome discrepancies.One—of particular interest in this analysis—is that in the latter partof the period considered here, the decline in the rate of change offertility was somewhat greater than one would have expected on thebasis of the two factors so far discussed. One possible explanation,suggested in several sources and consistent with the emphasis here oncompositional changes in the population, is an abrupt decline in theproportion of foreign-born women in the prime reproductive ages whocame from the high-fertility countries of southern and eastern Europe.There is substantial evidence that female immigrants from this areatypically had significantly higher fertility than contemporaneous im-

AND OF RATIO OF MALESAGED 20—34 TO FEMALES

1875—1930Per cent

2

(nc4le —)

- 120

100

go

80

70

60

50

40

0Females 20-34 —

emales 20-44(scale—)

15-19to

—10


migrants from northern and western Europe [82, p. 108; 189, pp. 4,10; 2111. Clearly, a sudden drop in the share of young foreign-bornwomen from this source would tend to depress fertility.

Direct evidence to test this proposition is not available sinceduring the period concerned the census did not regularly publish agedetail for the foreign born by country of origin. However, it seemspossible to form a rough impression of the validity of the argument.In the period 1890—1915, about two-thirds of all female immigrantscame from southern and eastern Europe; in 1915—30, about one-third.I have attempted to estimate, therefore, for foreign-born women aged20—84 at each of several dates, the proportion who had immigratedbetween 1890 and 1915, the peak period of the "new immigration." 17The results are as follows: 1900 45, 1910 = 82, 1920 = 86, 1930 =48. The figures clearly suggest a drastic decline during the 1920's inthe share of young foreign-born women accounted for by the newimmigration,'8 and thus appear consistent with the suggestion thatthe decline in the rate of change of foreign-born fertility during thisdecade, attributable in part to the demographic shifts previously noted,was aggravated by this factor.

Rural White FertilityThe explanation investigated here for Kuznets cycles in rural

fertility is a simple one; namely, that the rate of change of rural fertilityvaries directly with that in the economic condition of the farm popula-tion in family-building ages, approximated here by real farm incomeper head of the farm population (or labor force) as a whole. If therate of growth of real farm income per head drops off, the rate ofchange of farm fertifity would be expected to decline (algebraically).The converse is true if the rate of farm income growth increases.

17 The technique for 1930, for example, was to compare the number of survivorsfrom the group of foreign-born women aged 5—19 in 1915, estimated by appro-priate survival rates from [111, p. 23], with the number aged 20—34 enumeratedin 1930.

18 Thompson and Wheipton draw an opposite conclusion, namely, that the shareaccounted for by the new immigration rose slightly during the decade and thuscould not have contributed to the fertility decline [184, pp. 271—272]. The pro-cedure they use to infer the share of the new immigration, however, rests primarilyon figures for foreign-born women of all ages, and fails to take account of the factthat the major shift in national origins of immigration in the 1920's particularlyaffected the younger foreign-born age groups, those central to the explanation offertility.

The analysis comp1925—29 based on obserploying averages at quin58. In the first part of trural white rather than xiestimates for the former'present purpose and the'tively unimportant forei'fairly small. This seriesgross farm income perfor the latter allow for ifertility series. In the sethe birth rate for theconverted to five-year adecade, are comparedpopulation, again withformer.19 Both the quincdeflated by an index ch4ers. The series arepercentage change, our

By and large, as thiconsistent with the hypcgiven the shortcomingsany monocausal explanaincome per head or peby swings in the ratejoining scales willsubstantially greater thas suggesting an elasticconsistent with the find

If this reasoning ifertility change in

19 The shift to the farm Ibut more fundamentally toand farm income becomepopulation grows.

20 Cf. the studies of Ga[73, 161], Dudley Kirk (1[145a, pp. 237—255].


f82, p. 108; 189, PP. 4,

re of young foreign-borness fertility.n is not available sinceiiot regularly publish agerigin. However, it seems,alidity of the argument.,f all female immigrants1915—SO, about one-third.oreign-born women agedon who had immigratedhe "new immigration."= 82, 1920 = 86, 1930 =me during the 1920's in

for by the newpith the suggestion that,orn fertility during thisc shifts previously noted,

Xuznets cycles in ruralf change of rural fertilitytion of the farm popula-

by real farm incomeas a whole. If the

I drops off, the rate ofdecline (algebraically).ne growth increases.

the number of survivors1915, estimated by appro-

ber aged 20—34 enumerated

sion, namely, that the shareLuring the decade and thus14, pp. 271—272]. The pro-on, however, rests primarilyto take account of the factin the 1920's particularly

to the explanation of

The analysis comprises two parts, one for 1885—89 through1925—29 based on observations at decennial intervals; and one, em-ploying averages at quinquennial intervals, for 1920—24 through 1954—58. In the first part of the analysis, I use fertility data for the totalrural white rather than native rural white population, since the earlierestimates for the former are probably somewhat more reliable for thepresent purpose and the bias introduced by the inclusion of the rela-tively unimportant foreign-born group in the rural total is probablyfairly small. This series is compared with five-year averages of realgross farm income per person engaged in farming. The dates chosenfor the latter allow for a lead of one to one and a half years over thefertility series. In the second part of the analysis, annual estimates ofthe birth rate for the total farm population (white plus nonwhite),converted to five-year averages for the first and second half of eachdecade, are compared with real net farm income per head of farmpopulation, again with allowance for a lead of the latter over theformer.19 Both the quinquennial and decennial farm income series aredeflated by an index chosen to approximate the cost of living to farm-ers. The series are plotted in the upper panel of Figure 28, and thepercentage change, our particular interest, in the lower.

By and large, as the lower panel shows, the data seem reasonablyconsistent with the hypothesis—at least as consistent as one might hopegiven the shortcomings of the data and the inevitable limitations ofany monocausal explanation. Swings in the rate of growth of real farmincome per head or per worker appear to be matched fairly closelyby swings in the rate of growth of rural fertility. Reference to the ad-joining scales will show that the magnitude of the income swings issubstantially greater than that of fertility. This might be interpretedas suggesting an elasticity noticeably under one, a result which seemsconsistent with the findings of similar business cycle analyses.20

If this reasoning is accepted, then the historical course of ruralfertility change in this century would be conceived as reflecting in

19 The shift to the farm birth rate series is due in part to statistical convenience,but more fundamentally to the fact that the connection between "rural" fertilityand farm income becomes progressively more tenuous as the rural nonfarmpopulation grows.

20 Cf. the studies of Gary S. Becker [170, pp. 209—231], Dorothy S. Thomas[73, 161], Dudley Kirk [169, pp. 84-85; 170, pp. 241—257], and Morris Silver[145a, pp. 237—2551.

- "..— — -——— — -

significant measure thecharacterized farm inccfarm fertility in the 11the drastic setback tolowing World War I,

220 self was substantially200 I areas would be explaf

- income growth in thewith the war and postof growth of fertility:'

120meant a leveling off,farm boom and subst

too suggest that the advei90 has been somewhat80 possible reasons for th

the increased significa0 farm families [152, pp

30 tion of the "farm" pop:agriculture [198]; but

From what has. befor the total white poças a result of major flurural white componel'appear to have been cigration on the age, stborn, and of majorcondition of the farmeralizations based on(or of the entire poputhe total white) woul

Consider, for exagoing analysis is correoccurred in this decalexceptional circumstscomposition of the foimmigration of Wonstrictions, and, second

-

T98 / II: FERTILITY ANALYSES

FIGURE 28

LEVEL AND RATE OF CHANGE OF RURAL WHITE FERTILITY RATIO (ErR.) ANDREAL GROSS FARM INCOME PER ENGAGED 1885—1929; AND OF FARMismm RATE (F.B.R.) AND REAL NET FARM INCOME PER HEAl) 1920—58

80110 scales

-I

Y8 / N-(scale —I,

I1II

Yw I —

/II//•1 , —

I I I I I I

1.000

900

800C

10

500

400

-.,'

2054-58

1885- 95- 05- 15- 20- 30- 40- 50-89 99 09 19 24 34 44 54

StO

lg

1885-89 95-99 05-09 15-19 20-24 30-34 40-44 50-54to to to to to to to to

95-99 05-09 15-19 25-29 25-29 35-39 45-49 54-58

Sot.mcE; Tables C-9 and C-10.


significant measure the pattern of major surge and relapse which hascharacterized farm income growth. The accelerated rate of decline offarm fertility in the 1920's and early 1930's would be attributed tothe drastic setback to the growth of farm income in the period fol-lowing World War I, a decline so great that the absolute level it-self was substantially reduced. The subsequent baby boom in ruralareas would be explained by the corresponding resurgence in farmincome growth in the late 1930's and 1940's associated particularlywith the war and postwar booms. And finally, the decline in the rateof growth of fertility in the 1950's, which in terms of absolute levelmeant a leveling off, would be explained by the tapering off of thefarm boom and substantial drop in farm-income growth. The datasuggest that the adverse effect on fertility in this most recent periodhas been somewhat less than might have been expected. A number ofpossible reasons for this come to mind, such as compositional changes,the increased significance of nonfarm sources in the total income offarm families [152, pp. 48—49], and the progressive rise in the propor-tion of the "farm" population (1950 census definition) not engaged inagriculture [198]; but it is not possible to pursue these questions here.

From what has been said, it should be clear that the fertility trendfor the total white population has been subject to substantial variationas a result of major fluctuations in the fertility of the foreign-born andrural white components. The fluctuations for these groups in turnappear to have been caused by the impact of the rise and fall of immi-gration on the age, sex, and nationality composition of the foreign-born, and of major swings in agricultural conditions on the economiccondition of the farm population. It would seem to follow that gen-eralizations based on the fertility record of the total white population(or of the entire population, whose behavior is of course dominated bythe total white) would be extremely hazardous.

Consider, for example, the experience of the 1920's. If the fore-going analysis is correct, the striking decline in total white fertility thatoccurred in this decade was caused largely by the conjuncture of twoexceptional circumstances—namely, major shifts in the demographiccomposition of the foreign-born population arising from the effect onimmigration of World War I and the subsequent imposition of re-strictions, and, second, a major slump in agricultural conditions. When

•1 - -_, 9'T '

RTILITY RATIo (R.F.R.) AND1885—1929; AND OF FARME PER 1920—58

180

160

140'

120

100

90

80

70

30

-..n

a


added to the continuous depressing influence of the rural-urban shift,these circumstances created a decline in white fertility noticeablyout of line with previous experience. Knowing this, one is inclinedto view with some reserve statements such as that quoted previously,which cites the sharp fertility decline for the total population in theprosperous 1920's as a reason for discounting the effect of economicconditions on fertility.

It is nevertheless true that even urban native white fertility de-clined in this decade, though the decline of under 6 per cent for thisgroup is rather less impressive than the almost 20 per cent decline forthe white population as a whole. It is time, therefore, to see whatmight explain the fertility pattern for this group.

Urban Native White FertilityAs in the rural analysis, the aim here is to explore the relation

between Kuznets cycles in fertility and in the economic condition ofthe population of family-building ages. For the rural population, itseemed reasonable to assume that the economic experience of those infamily-building ages could be inferred from the income experience ofthe farm population as a whole. Such an assumption, however, doesnot seem plausible for the urban group, with its much more varieddistribution of industrial and occupational attachments. In the absenceof direct information on the situation of those in family-building ages,therefore, I have attempted to infer the state of the labor market foryoung persons from two indicators, conceived as reflecting respectivelythe demand and supply sides of the market. The first is the unem-ployment rate for the labor force as a whole. A low rate is taken asreflecting a generally favorable state of demand for labor, young andold; a high rate, an unfavorable situation. The second is the rate ofchange of the total white male population, aged 20—29, taken as acrude index of the rate of entry of young persons into the labor market.Other things equal, a decrease in the rate of entry would make for afavorable labor market for young persons because of their scarcity;an increase, an unfavorable market. Thus the hypothesis is that therate of change of urban native white fertility varies directly with thatof aggregate labor demand (read "inverted unemployment rate") and

inversely with that of th(read "rate of change of

An example may ciaing a Kuznets-cycle expwould increase and, oth(pect this to lead, througitions, to a favorable res:

• encouraging marriage aof free immigration, thiwould also provoke an Irate of additions to the

• tendency toward tighttstimulus to fertility ofimmigrants are typicall)which we are intereste(tendency toward an adcreased rate of growthtraction would be modeThus Kuznets cycles in tto be compensated by sket owing to immigrati(fertility would be count

Figure 29 presentsshows the levels of thefacilitate inferences froi

21 Although the view thatreflect changes in aggregatiis not essential to the analygeneral unemployment rateditions, i.e., the net outcoin the rate of entry as inof the situation forperiod covered here a riseployment seems consistentcols. 2, 3).

22 Some may note a siranalysis emphasizing the anative population [212, 21trend, whereas the presenttakes account of the stimulwhich encourage a rise in I

-


of the rural-urban shift,vhite fertility noticeablying this, one is inclined

that quoted previously,total population in thethe effect of economic

native white fertility de-under 8 per cent for thisit 20 per cent decline for

therefore, to see whatpup.

to explore the relatione economic condition ofthe rural population, itc experience of those ine income experience of

umption, however, doesits much more varied

chments. In the absencein family-building ages,of the labor market fors reflecting respectivelyThe first is the unem-A low rate is taken as

vi for labor, young andse second is the rate ofraged 20—29, taken as a

s into the labor market.ntry would make for aause of their scarcity;hypothesis is that the

aries directly with that• employment rate") and

inversely with that of the rate of labor market entry of young persons(read "rate of change of white male population, aged 20-29") •2t

An example may clarify the reasoning. If the economy is experienc-ing a Kuznets-cycle expansion, the rate of growth of labor demandwould increase and, other things remaining unchanged, one would ex-pect this to lead, through its effect on income and employment condi-tions, to a favorable response in fertility of the native population byencouraging marriage and childbearing. However, under conditionsof free immigration, the increased rate of growth of labor demandwould also provoke an influx of immigrants. The resulting rise in therate of additions to the labor market would tend to counteract thetendency toward tightening and thus offset in some measure thestimulus to fertility of the native born. Note, in this connection, thatimmigrants are typically concentrated in exactly those age groups inwhich we are interested for the analysis of fertility. Conversely thetendency toward an adverse impact on native-born fertility of a de-creased rate of growth of labor demand during a Kuznets-cycle con-traction would be moderated by a decrease in the rate of immigration.Thus Kuznets cycles in the rate of change of labor demand would tendto be compensated by swings in the rate of entry into the labor mar-ket owing to immigration, and the consequent impact on native-bornfertility would be counteracted in some degree.22

Figure 29 presents the relevant series; as before, the upper panelshows the levels of the variables, the lower, their rates of change. Tofacilitate inferences from the graph, the curve for each of the explana-

21 Although the view that variations in the general unemployment rate primarilyreflect changes in aggregate demand seems most consistent with formal theory, itis not essential to the analysis. Alternatively, one might think of movements in thegeneral unemployment rate as indicating the average course of employment con-ditions, i.e., the net outcome of aggregate demand and supply, and changesin the rate of entry as indicating. variations in the deviation from the averageof the situation for young persons. However, the fact that for most of theperiod covered here a rise in the rate of entry accompanied a reduction in unem-ployment seems consistent with the emphasis on aggregate demand (Table C-li,cols. 2, 3).

22 Some may note a similarity between this reasoning and Francis Walker'sanalysis emphasizing the adverse influence of immigration on the fertility of thenative population [212, 213]. Walker, however, was concerned with the primarytrend, whereas the present ana'ysis refers only to Kuznets cycles, and in additiontakes account of the stimulating influence to native fertility of the very conditionswhich encourage a rise in immigration.

FICUER 29LEVEL AND RATE OF CHANGE OF URBAN NATIVE WHITE FERTILITY RATIO

UNEMPLOYMENT RATE OF CIVILIAN LABOR FORCE (u), AND RATEOF CHANGE OF TOTAL WHITE MALE POPULATION ACED 20—29 (s), 1885—1958

-w

Table C-il.

or,, U,

a

a

r

tory variables has been1would be expected to caiother things remaining t

If we first considethrough 1935-39, the maof the two explanatoryrate of growth of aggmoves in a way favorab]young persons intoversely, and vice versa:suppiy conditions refiecthe situation describedment reflects primarilyFor example, the increapared with that in thethe rate of increase of rturn to a correspondingin the century.

So far as directionsthis period are concerrregarding the expectedplus in one is accompafertility curve fails toplanatory series. Ratherbeginning through therelatively small declinesistent part of this pictions being offset bywere to smooth out flua simple two-item moployment rate, would sthe fertility curve. An.produce the same effec

Still more intriguii39. In this period, thepattern of rise and falthan previously. In stria Total white, 1925—29 on.

THE AMERICAN BABY BOOM / 103WEITE FERTILITY RATIO

OR FORCE (u), Aiw tory variables has been plotted inverted so that an upward movementED 20—29 (s), 1885—1958 would be expected to cause an upward movement in the fertility curve,

other things remaining unchanged.If we first consider variations in the decade rates of change

through 1935—39, the most interesting feature is the inverse movementsof the two explanatory series. As the lower panel shows, whenever therate of growth of aggregate labor demand (the lower solid line)moves in a way favorable to fertility, the change in the rate of entry ofyoung persons into the labor market (the broken line) moves ad-versely, and vice versa. In the early part of the period the swing in

osupply conditions reflects chiefly movements in immigration—exactlythe situation described in the example above. Later, the supply move-

10ment reflects primarily variations arising from demographic sources.For example, the increase in the decade from 1915—19 to 1925—29 corn-pared with that in the decade preceding reflects an exceptional rise in

20 . .

the rate of mcrease of native white males aged 20—29, which traces inturn to a corresponding movement in the total white birth rate earlierin the century.

So far as directions of movement of the explanatory series duringthis period are concerned, therefore, they carry no clear implicationregarding the expected behavior of the rate of change of fertility—aplus in one is accompanied by a minus in the other. And, indeed, thefertility curve fails to exhibit the fluctuations of either of the two ex-planatory series. Rather, one finds simply one extended swing from thebeginning through the period 1925—29 to 1935—39. The 1920's, with arelatively small decline in the rate of change of fertility, form a con-sistent part of this picture, a favorable movement in demand condi-tions being offset by an adverse one in supply. Interestingly, if onewere to smooth out fluctuations in the two explanatory curves by, say,a simple two-item moving average, both, and particularly the unem-ployment rate, would show an extended swing rather similar to that ofthe fertility curve. An average of the two explanatory curves wouldproduce the same effect.

Still more intriguing is the behavior of the three series after 1935—39. In this period, the rate of change of labor demand continues itspattern of rise and fall, with a swing of noticeably greater amplitude

on. than previously. In striking contrast to the preceding pattern, however,

- -


the change in the rate of entry into the labor market levels off insteadof fluctuating inversely. And, for the first time, the rate of change offertility exhibits a Kuznets-cycle movement, reproducing with re-markable similarity the fluctuation in the rate of change of labor de-mand. The inference suggested by these movements seems clear. Withimmigration restricted and without a surge in the rate of labor mar-ket entry from the native-born population caused by demographicprocesses, the favorable impact of a swing in the rate of growth ofdemand—itself much larger than heretofore—was felt with muchgreater force by the young native whites in the labor market. As aresult, the rate of change of fertility of this group reproduced theswing in labor demand in significant measure for the first time.

If one considers magnitudes of the variables rather than simplyrates of change, the argument seems reasonably well borne out, thoughthe correlation is not perfect. In Table 2, each of the seven observa-tions on the rate of change in fertility is classified according to theaccompanying values of the rate of change in the unemployment rateand in the percentage change in white males aged 20—29. One findsthat, holding the change in rate of entry into the labor market con-stant (that is, examining each row in the table separately), the rateof change of fertility varies directly with the rate of change in demand(inversely with the rate of change in the unemployment rate). Con-versely, holding demand conditions constant (examining each column

TABLE 2. OBSERVATIONS ON PERCENTAGE BATE OF CHANGEPER DECADE IN URBAN NATIVE WHITE FERTILITY, CLASSIFIEDBY CONCURRENT CHANGE PER DECADE IN PERCENTAGE OFLABOR FORCE UNEMPLOYED AND IN PERCENTAGE RATE OFCHANGE OF TOTAL WHITE MALES, AGED 20-29, 1885-1958

Change per Decade in PercentageChange per Decade in Per-centage Rate of Change of

Total White Males, Aged 20—29

Unemployed (percentage points)

—2 to(percentage points) —16 —8 +2 +7 +14

+8 to +10 +2 —6—2 to —5 +42 +18 —8 —22

—18 0

SOURCE: Table C-li.

• = •- - __1

separately), there is avary inversely with themarket, though in thisentries being out of ordmanly reflects a fundaminadequate approximaticbuilding age provided 1possible to say.

A comprehensive xperience of young persoiingly out of reach. Yetto assemble supports thmarket situation of yolrecent years. Consider t

1. In the 1940's,much more rapidly thaiat about the same or app. 6—7, and No. 35, Jashighly concentrated inthey must haveThe very fragmentarydevelopment in the 192(

2. The shift of youlceeded at a much greatdecades. In 1940, 17 petions were in the threeand kindred workers; ITand craftsmen, foremerof this same group of ran improvement of 24provement for the coho;

2S In the examples cited1950's is consistent with this, the abrupt break with ifavorable situation for yawThe movement from the 11of the process. It is likely Ithis pattern would be still

THE AMEBICAN BABY BOOM / 105

market levels off insteadne, the rate of change ofLt, reproducing with re-

of change of labor de-ements seems clear. With'n the rate of labor mar-caused by demographicin the rate of growth ofe—was felt with muchthe labor market. As a

s group reproduced thefor the first time.

ibles rather than simplyy well borne out, though

of the seven observa-tssifled according to thethe unemployment rate

s aged 20—29. One findso the labor market con-)le separately), the rateite of change in demand

rate). Con-examining each column

RATE OF CHANGERTILITY, CLASSIFIED

PERCENTAGE OF2ENTAGE RATE OF20—29, 1885—1958

r Decade in Percentageed (percentage points)

—2 to+2 +7 +14

—6+18 —8 —22

I

separately), there is a tendency for the rate of change of fertility tovary inversely with the change in the rate of entry into the labormarket, though in this case there is one inconsistency (the +18 and 0entries being out of order vertically). Whether this discrepancy pri-marily reflects a fundamental deficiency in the analytical scheme or aninadequate approximation to the economic condition of those of family-building age provided by the explanatory series used here, it is notpossible to say.

A comprehensive measure of the income and employment ex-perience of young persons for the period covered here remains tantaliz-ingly out of reach. Yet such additional evidence as I have been able

• to assemble supports the view that the income experience and labormarket situation of young persons were exceptionally favorable inrecent years. Consider the following: 23

1. In the 1940's, earnings in the lower-income occupations rosemuch more rapidly than those in the higher, and then, in the 1950's,at about the same or a slightly lower rate [180, No. 33, Jan. 15, 1960,pp. 6—7, and No. 85, Jan. 5, 1961, p. 521. Since young people are morehighly concentrated in lower-income occupations than older people,they must have particularly benefited from the movement of the 1940's.The very fragmentary evidence available suggests no correspondingdevelopment in the 1920's.

2. The shift of young persons into higher-earning ocãupations pro-ceeded at a much greater rate in the 1940's than in the two precedingdecades. In 1940, 17 per cent of males aged 15—24 in nonfarm occupa-tions were in the three highest income classes (professional, technical,and kindred workers; managers, officials, and proprietors, except farm;and craftsmen, foremen, and kindred workers). By 1950, 41 per centof this same group of males (now aged 25—34) were in these classes,an improvement of 24 percentage points. From 1920 to 1930, the im-provement for the cohort moving through the same ages was 17 points,

23 In the examples cited, the typical movement from the 1930's through the1950's is consistent with the pattern shown by the rate of change of fertility—thatis, the abrupt break with past experience, in a direction reflecting a particularlyfavorable situation for young persons, occurs between the 1930's and the 1940's.The movement from the 1940's to the 1950's suggests a slowing or even reversalof the process. It is likely that between the first and second halves of the 1950'sthis pattern would be still more apparent (cf Chapter 5).


and from 1930 to 1940 it was 12 points. Corresponding figures for thecohorts aging 25—34 to 35—44 in the three successive decades are 7, 4,and 14 points. Other things being equal, this more rapid shift tohigher-income occupations points to a significantly higher rate of in-come growth for young persons in the 1940's than in the two preceding

8. Expansion of government transfer payments provided a newbulwark to income in the 1940's and 1950's, especially in the form ofveterans benefits and unemployment compensation for younger persons.

4. Labor force participation rates in the 1940's showed a markedbreak with previous trends in a manner strongly suggesting a shortageof young workers. The sharp downtrend in participation of white malesaged 14—19 which had prevailed since 1900 was completely reversed.A similar movement appears even to have characterized those aged10—13 [111, pp. 364—367]. The long-term rise in labor force participa-tion of older women was greatly accelerated because jobs that wouldordinarily have been filled by young persons were left open. Andwhile, for young women as a whole, labor force participation declinedslightly as a larger proportion married and had children, the rates forwives, even those with preschool-age children, rose substantially.Finally, while it is not possible to cite figures on the long-term trend,part-time employment rose substantially after 1940, and it seems likelythat this too stemmed at least in part from a shortage of young work-ers. In the 1950's the rise in labor force participation of older womencontinued virtually unabated, but the rate for those aged 14—19 re-sumed its long-term decline.25

5. Since 1940, home ownership among young persons has risen tolevels markedly higher than had previously prevailed. The followingfigures for nonfarm household heads show, for each age group, thepercentage of dwelling units which were owner-occupied at eachdate: 26

24 The figures for 1930—50 are computed from [94, Appendix Table 1]; for 1920,from unpublished estimates comparable to [94] kindly provided by W. Lee Han-sen. Data for armed forces as reported in the census were included with the1940 and 1950 figures.

25 The evidence cited in this paragraph is from the excellent census monographby Gertrude Bancroft [12, pp. 29—31, 58, 77—82, and Ch. 4). See also Part IIIbelow.

26 The data through 1940 are from the census reports; for 1949 and 1959, from

Age 189015—24 1425—34 2435—44 35

There is a marked ad1940, part of which muavailability but to a siwhich encouraged takiz

6. Finally, there axrecent study [82] has sIthe significant decline1940, age at marriagemedian for all femalesdecade, a period one-I[186, Series A-229]. Inafter marriage. These Ibearing, rather thanaccounted for most of t365_371].27 The centraobvious. It would beincome and ernployme

CONCLUSIONS A

The most striking feat,most urgently for expitorical experience. Hov

[91, p. 1107, Suppi. Tab!were adjusted to 15—24 aown their own homes.) TIhold only, which biasesother dates.

27 The draft law policyriage and childbearing, btthe family beyond the Sibaby boom of the typeaverage number of childadevelopment is of coursethe exceptional labor maspouse will take this form


Age 1890 1900 1930 1940 1949 195914 10 11 12 21 1624 21 28 22 85 4235 34 44 37 58 63

There is a marked advance in the situation of young persons after1940, part of which must be due not only to a great increase in creditavailability but to a substantially improved income position as wellwhich encouraged taking on long-term commitments.

6. Finally, there are the characteristics of the baby boom itself. Arecent study [82] has shown that a major factor in the boom has beenthe significant decline since 1940 in age at marriage. From 1890 to1940, age at marriage drifted irregularly downward, the decline in themedian for all females amounting to oniy one-half year. In the nextdecade, a period one-fifth as long, the reduction was twice as great[186, Series A-229]. In addition, wives have had children much soonerafter marriage. These two factors, earlier marriage and earlier child-bearing, rather than mothers having substantially more children,accounted for most of the rise in the fertility rate through 1954 [82, pp.365_371].27 The central role of young families in the baby boom isobvious. It would be difficult indeed to account for this unless theirincome and employment experience had been exceptionally good.

CONCLUSIONS AND POSSIBLE IMPLICATIONSThe most striking feature of the baby boom—and thus the one callingmost urgently for explanation—is the apparent abrupt break with his-torical experience. However, reconciliation of present and past becomes

15—2425—3435—44

esponding figures for theccessive decades are 7, 4,this more rapid shift tocantly higher rate of in-han in the two preceding

yments provided a newespecially in the form ofLtlon for younger persons.1940's showed a marked

suggesting a shortageticipation of white malesvas completely reversed.

those agedin labor force participa-because jobs that wouldis were left open. Ande participation declinedd children, the rates forren, rose substantially.on the long-term trend,

1940, and it seems likelyhortage of young work-ipation of older womenr those aged 14—19 re-

ing persons has risen torevailed. The followingor each age group, thewuer-occupied at each

ppendix Table 1]; for 1920,provided by W. Lee Han-'s were included with the

excellent census monographCh. 4). See also Part III

s; for 1949 and 1959, from

[91, p. 1107, Suppl. Table 1]. (Data for those aged 18—24 from the latter sourcewere adjusted to 15—24 on ±e assumption that no heads of households under 18own their own homes.) The 1930 and 1940 estimates are for male heads of house-hold only, which biases them slightly upward compared to the figures for theother dates.

27 The draft law policy of deferring fathers doubtless encouraged earlier mar-riage and childbearing, but without an income situation that favored expansion ofthe family beyond the first child, it is doubtful that it could have produced ababy boom of the type experienced. There is now reliable evidence that theaverage number of children per mother has also risen in the postwar period. Thisdevelopment is of course consistent with the analysis presented here. The longerthe exceptional labor market situation prevails, the more likely the fertility re-sponse will take this form in addition to earlier marriage and earlier childbearing.


process was further abettecditions, which evoked a sirural white population.

In conclusion, Some offor the past and future ma

With regard to thewhile Kuznets cycles in thphenomenon in our historyfrom immigration to fertilitbears any logical connectiopresent analysis is that mias we permitted freeresponse to swings in labo:this country acted as anative white population.the urban native white poswing in labor demand wian unparalleled responsepopulation growth.

The implications of thof fertility change are inthe typical demographictomary emphasis ofthe past would suggest aresumption of the primarthe present analysis, how

easier when one recognizes that even before the 1940's the historicalrecord was characterized by fluctuations of significant magnitude andduration, and that the record for the total white population is a com-posite of the varying experience of several component groups, subjectin part to quite different influences. Major swings in agricultural con-ditions, on the one hand, and Kuznets cycles in nonagricultural activitywith accompanying immigration fluctuations, on the other—each withtheir peculiar historical timing—gave rise to distinctive fertility re-sponses on the part of the rural white, foreign-born white, and urbannative white populations. When one unravels these differing strandsof experience and considers their underlying influences, the impressionemerges that the recent fertility behavior of the urban native whitepopulation, the group of central significance for explanation of thebaby boom, is not as inconsistent with its earlier character as washeretofore believed. In the first three decades of the century, thefertility of this group, instead of exhibiting a declining trend, showedreasonable stability. And in the recent period the effect on the labormarket of a Kuznets-cycle expansion—an expansion stronger, accordingto our data, than any preceding ones considered here—was for thefirst time not accompanied by an offsetting rise in the rate of labor-market entry due to a significant increase in either immigration or thenative-born population in young working ages. The unprecedentedconcurrence of these three circumstances—a Kuznets-cycle expansionin the economy, restricted immigration, and a low rate of labor forceentry from the native population resulting from demographic processes—created an exceptional job market for those in family-building agesand as a result drastically accelerated the founding of families.28 This

28 With regard to the causes of the exceptional labor market for young personsin the 1940's and 1950's, the present chapter emphasizes quantitative scarcity tothe exclusion of relative quality. The following figures on median school yearscompleted by young and middle-aged males at various dates may partially rightthe balance (see also Figure 19):

Age at 1970Specified Date 1920 1930 1940 1950 1960 (projected)(1) 25—29 8.4 8.7 10.1 12.0 12.3 12.5(2) 45—54 8.1 8.2 8.4 8.7 10.0 12.0(3) (1)—(2) 0.3 0.5 1.7 3.3 2.3 0.5

Note the immense gain in the educational advantage of young over middle-agedworkers in the 1940's, a change which sharply improved their competitive position

at just the time that labor denming from the abrupt advanoccurred in the 1920's and esg236, 2381 and [182, pp. 6—7].as those reported by the cornon educational attainment wer

The sequence of change inpronounced convergence in iiextent the change in the size ding income differentials by ageber and quality of young worli

29 Clearly the presenttrend itself in terms of the dmight prove fruitful.

1.

re the 1940's the historicalsignificant magnitude and

white population is a corn-component groups, subjectswings in agricultural con-in nonagricultural activity

s, on the other—each withto distinctive fertility re-ign-bom white, and urban'els these differing strands

influences, the impressionof the urban native whitece for explanation of the

earlier character as was

athe the labor

ansion stronger, accordingsidered here—was for therise in the rate of labor-either immigration or theages. The unprecedented

a Kuznets-cycle expansionI a low rate of labor forceom demographic processesse in family-building ages)unding of families.28 This

thor market for young personshasizes quantitative scarcity togures on median school yearsrious dates may partially right

50 1960

LO 12.31.7 10.01.3 2.3

1970(projected)

12.512.00.5

ge of young over middle-agedved their competitive position


process was further abetted by a concurrent boom in agricultural con-ditions, which evoked a similar fertility response on the part of therural white population.

In conclusion, some of the implications of the preceding analysisfor the past and future may be set forth.

With regard to the past, it was noted earlier in the discussion thatwhile Kuznets cycles in the rate of population growth are not a newphenomenon in our history, the shift in the source of these movementsfrom immigration to fertility raises a question whether the recent cyclebears any logical connection to its predecessors. The implication of thepresent analysis is that indeed such a connection does exist. As longas we permitted free immigration, the rise and fall of immigration inresponse to swings in labor demand associated with Kuznets cycles inthis country acted as a buffer to moderate the impact on the urbannative white population. With the restriction of immigration, however,the urban native white population felt the impact of a Kuznets-cycleswing in labor demand with unprecedented force, and the result wasan unparalleled response in fertility and thus again in the rate ofpopulation growth.

The implications of the present analysis for the longer-term futureof fertility change are in contrast with that likely to be suggested bythe typical demographic discussion of our fertility history. The cus-tomary emphasis of demographers on the long-term secular decline inthe past would suggest a view of the current fertility decline as aresumption of the primary trend.29 The interpretation suggested bythe present analysis, however, would be that for the group whoseat just the time that labor demand was booming. The timing is fortuitous, stem-ming from the abrupt advance in the diffusion of high-school education thatoccurred in the 1920's and especially the 1930's. (The figures are from [178, pp.236, 238] and [182, pp. 6—7]. The 1920 and 1930 values were assumed the sameas those reported by the corresponding cohorts in 1940, the first time that dataon educational attainment were collected.)

The sequence of change in the educational differentials calls to mind the recentpronounced convergence in income distribution by size. One wonders to whatextent the change in the size distribution in the past forty years may reflect chang-ing income differentials by age associated with variations in both the relative num-ber and quality of young workers.

29 Clearly the present analysis suggests that a reexamination of the primarytrend itself in terms of the differing patterns of the groups distinguished heremight prove fruitful.

- - I -- U'


experience is of central significance for the future, the urban nativewhite population, the nature of the primary trend in this century—whether upward or downward—is not readily apparent, and the recentbehavior of this group may be largely explained as a Kuznets-cyclephenomenon. If this is correct, then substantial fertility variation, up CHAPTER 5 / T For down, may Qçcur again over the longer run. FE R T I L I T Y D

PROJECTED F

The postwar peak of fertitotal fertility rate hasfew years quite substantposable income, a comncorrelations with fertility;slightly lower than previhave had little to do widoubts about the explanaWhile a thorough investhere, it is possible to dfactors have changed indine, and to considerAs in Chapter 4, the focuNo attempt is made tonativity or other componcreasing homogeneity ofency. The more limited tithe present analysis largforce surveys of the pastdata heretofore unavailagraphic behavior of cornwith our interest in theyear fluctuations, these ca three-year moving aveon comparisons betweencurrent fertility decline.of the series become av

chapter 4 / the american baby boom in historical

Documents