Economics in Times of Crisis

Cliometric Sessions at 1990 ASSA Meetings--December 28, 8:00 AM



William H. Phillips
University of South Carolina

This paper examines how the rapid growth of demand at the beginning of World War II disrupted the family-economy basis of a southern textile mill town in South Carolina. Just prior to the war, women textile hands were only a secondary labor force tied to their husband's job at the mill. As a result, older, more experienced female workers had no market power to push for promotions and higher earnings. Once the wartime economy soaked up the surplus labor of the preceding depression, female mill operatives began to have economic opportunities independent of their spouse's employment. Therefore, one sees in 1942 a growing tendency for the mill to place more experienced female hands into higher- paying positions. An interesting question for future research will be to investigate whether this isolated mill town reverted back to its family economy basis once the war was over. In order to conserve space, this summary omits all figures, tables, and footnotes, which may be obtained by writing the author.

The Second World War stands out as a crucial watershed in Southern economic development. As documented by Gavin Wright in Old South, New South, wartime demands and government decisions about the administration of wage controls forced the South out of its labor market isolation. A new mobility and awareness of outside opportunities was created, and the region could never simply return to `normal' once hostilities ceased. In my own personal case, my newly-married father was ready to settle down as the principal of a school in a very small Alabama town when the war began. The army then inducted him in order to teach science and math to would-be Air Force officers in the state capital of Montgomery. My parents were forced much against their will to move only a hundred miles, but effectively into a foreign culture. By the time the war was over, they had absolutely no intention of ever moving back into their old world. And although this change enhanced the economic possibilities for the both of them, certainly this event meant the most to my mother's future career opportunities.

The purpose of this paper will be to examine one instance of this labor market revolution for Southern women and Southern manufacturing at the beginning of the war. The Courtenay Textile Mill of Newry, South Carolina, was an isolated company town built in 1892, and had been a typical part of the Piedmont cotton mill culture of the early twentieth century. Payroll records kept from 1937 on, as well as a personnel file maintained from 1942 to 1950, allow a unique opportunity to survey the structure of its male and female jobs and earnings from the late depression through the post-war era. This particular study will be based on estimated age-earnings profiles for both male and female Courtenay employees from early 1937 to the last quarter of 1942.


The cotton textile industry labor market known as the `family economy' has been the subject of intense interest by both historians and economists. In this section, a simple model of the family economy for the mature cotton textile industry of the Depression South will be constructed. Moreover, the impact of minimum wage legislation must be considered. From this, certain implications about the expected shape of an age- experience earnings profile will be derived. The subsequent section of the paper will then test the model with Newry Mill payroll data from 1937 to 1942.

A key advantage to the mills afforded by the family economy labor system was that it allowed them to `hoard' labor against fluctuations in demand. Despite the development by the Depression era of a fairly experienced cotton mill labor pool all throughout the Piedmont, the search for new labor from outside the town to fill positions was costly. This was particularly the case in a competitive environment of many small firms. Market demand was very sensitive to the business cycle and completely outside the control of businesses like the Newry Mill. When orders suddenly poured in, the mills needed additional labor immediately. Yet when the inevitable downturn came, they wanted to both lay off excess workers, yet know that this labor was available at a moment's notice. Economists have long speculated that many industries hoard labor in cyclical downturns by not laying off as many workers as they could really do without. The firms `invest' in a stand-by workforce by continuing to pay them. In this fashion, the whole mill village system and its family basis can be interpreted for the mature textile industry as an investment designed to keep its carefully built up labor force from disappearing during layoffs.

From the very beginnings of the late nineteenth century Southern textile economy, the mills' labor force had been comprised of the primary family bread-winners, supplemented by a secondary family-labor pool that was available for work consistent with the needs of the mill and a particular family's situation. Furthermore the sex and age composition of this secondary labor pool changed with the evolution of the industry and its workers. The early days of the industry are filled with references to the `bucket-toters.' These were older farmers who had moved their families to the mills, but who were really past the age of easily learning the new skills demanded by the firm. Although this secondary labor group was an easy target for the progressive supporters of child-labor laws, their wives and children were in these early days much more likely to be able to adjust to the new industrial world. Moreover, as Gavin Wright has noted, this arrangement was particularly suited to farmers with a preponderance of female children, who had no opportunities in farming. The primary labor pools of the early years were wives and female offspring, with a secondary role played by the ex-farmers and their sons, who could move back and forth for a few years between textile and agricultural labor.

Child-labor, especially of the younger ages, also certainly played a role as a secondary labor-force for the mill. Cathy McHugh has shown how mill-supported schools acted as a screening device to attract more sober and stable families to the village. But another function noted by her is that of providing a central location where the children could be found when extra hands were needed. In this regard, it is interesting to note that once child-labor restrictions appeared inevitable, the mills pushed for accompanying compulsory education legislation. This quickly led to the necessity of state-supported education, which provided the mills with an opportunity to relieve themselves of some or most of their educational cost burden. A burden that now had less compensation if the school-children were not available for work.

Eventually, however, the role of supplementary labor fell to the women of the mature Southern textile industry. With the steady increase of male textile workers who had grown up in the villages, the bucket- toters disappeared. Moreover, with access to this primary labor pool supplemented by their wives and older children, employer resistance to child-labor laws lessened. Women, who had comprised 57.6 percent of the Southern textile labor force in 1880, held only 39.1 percent of the cotton mill jobs by 1930. This trend was only accentuated during the Great Depression, and increasingly a women's textile job came only as a package supplementing her husband's or father's job. As shown in Table 2 in the appendix to this paper, a matched payroll-personnel file sample for early 1937 at the Newry Mill indicates that women comprised only 21 percent of the mill labor force. Moreover, only 38 percent of the men in the sample working at the mill also had a wife working there, whereas almost 81 percent of the women had a husband working at Newry.

Figures 1(a) and 1(b) provide a demand and supply model of the family economy labor markets in the mature Southern textile industry of the Great Depression. As the primary labor force of the mill, male labor had the only realistic possibility of searching for a better job, however limited it might be in the face of high unemployment. In this case, older, more experienced labor is distinguished from younger, less experienced labor, in three ways. First, there is a relatively greater demand by firms for older labor (D versus DL), based on its greater productivity, or as a reward for its loyalty, reliability or seniority. Second, experienced labor S is in less elastic supply than inexperienced labor, SL, since an increase in demand will more quickly exhaust any local surplus of seasoned hands, necessitating faster growth in wages to attract such hands from outside areas. Finally, experienced labor will be in reduced supply relative to inexperienced labor, since such workers will have better economic opportunities to obtain jobs at other mills.

The female labor force, on the other hand, is now a supplementary, surplus labor force tied to their family's position at the mill. Older, experienced women operatives have no differential outside job opportunities over inexperienced women operatives. Other mills will only offer them possible employment as an adjunct to hiring their husbands. Moreover, in this depressed labor market, mills have many unemployed women operatives married to husbands working at their mills. Any increase in demand for labor can easily be filled by tapping this pool at current wages. The supply of both experienced and inexperienced labor is the same (SL) and is perfectly elastic.

The result of this labor market scenario is that whereas the male textile operatives should have a typically-shaped age-earnings profile, women hands should see no meaningful relationship between age, experience, or earnings. In Figure 1(a), the higher demand D for experienced male workers, combined with reduced supply S, implies a positive correlation between experience (and therefore age) and the expected wage rate. For women, however, the extra demand D for experienced labor will only be met out of the surplus labor pool at current wages. This would imply a perfectly flat age-earnings profile for women textile operatives in the mature and depressed industry.

It should be noted that this analysis is not saying that the mill family did not care whether other family members were hired at the mill, and what wages they were paid. Obviously they did, and certainly the expectation of a family that moved into the village was that in normal times, places would be found for all or most of them. What would not be expected once this male primary labor force had been established, was that the supplemental jobs held any chance for promotion or advancement. The young male hands would start off at minimum-wage jobs, and should they show some reasonable ability at them, they stood a definite chance of eventually working up to weaver, and more rarely, up to loom fixer or section hand in their middle age. Many males who stayed did not advance, but they were outnumbered by those that did.

Women operatives in this era, on the other hand, were put in the spinning room or on batteries, and generally stayed there. Although a few women would become highly-paid weavers, as can be seen in the `Early 1941' job ladder for Newry Mill women in Table 1, they were greatly outnumbered. Nor did age seem on the eve of the war to have much to do with the selection of women for the looms. Spinners in the early 1941 sample were, for example, older on average (35 to 31 years) than those women in the weaving room.

Finally, the minimum-wage initiatives of the NRA and later the Fair Labor Standards Act would also be expected to contribute to this supplementary labor status for women. First, the unemployment effects of the wage increases would swell the labor surplus that the mills could draw upon, particularly for the women more concentrated in the lower-paying jobs relative to men. Secondly, given the preference by the mills at this time for an adult male labor force, the floor to wages will induce a further substitution of lower-level male jobs for now equally-priced female jobs. The age-earnings profile is even more likely in this case to show no impact of age on earnings for women, particularly if the legislation is binding enough to force jobs staffed by younger women to be paid equally to jobs staffed primarily by older women.


The family economy model of the previous section was tested by estimating age-earnings profiles separately by sex for a sample of employees at the Newry Mill. Payroll data on average hourly earnings was obtained for the first week of 1937, 1938, 1939, and 1940 respectively. The first five weeks of average hourly earnings were obtained for 1941, while a surviving payroll summary book enabled the easy coding of average earnings for the entire year of 1942, broken down into quarters. Matches were then obtained between the payroll sample and personnel card information on sex, age, place of birth, and family information.

The log of average hourly earnings was then regressed by OLS on age, age-squared, birthplace, family variables, and the log of per pupil educational expenditures for the worker's county of birth. The mean values of the variables, and their definitions are presented in Table 2 in the appendix. Different specifications of the non-age variables in the model had no meaningful impact on the estimated age and age-squared coefficients. Moreover, results for early 1938, 1939, and 1940 were essentially identical to 1937 and 1941, as were the second and third quarters of 1942 to the first and fourth quarters. Therefore, Table 3 in the appendix only presents the estimated age-earnings coefficients for Early 1937, Early 1941, Early 1942 (first quarter) and Late 1942 (fourth quarter).

Figures 2 and 3 depict the expected age-earnings profiles by sex for Newry Mill operatives in Early 1937 and Early 1941. They were constructed from the coefficient estimates of Table 3 and the mean values of the non-age variables of Table 2, and represent the predicted hourly earnings at various ages for an operative possessing characteristics average to his or her sex. The estimated coefficients and profiles show that as expected, there was a statistically significant correlation between male earnings, age and age-squared for both 1937 and 1941. The average male operative advanced to higher-paying positions until approximately forty years of age, after which earnings stagnated or declined. Therefore, these results are consistent with the male family economy labor market of Figure 1(a). Experienced male textile workers commanded a premium in earnings based on additional demand for their services and greater outside job opportunities.

Women at the Newry Mill, however, possessed no statistically significant advancement of earnings with age from 1937 to 1941. The signs of the estimated age coefficients, moreover, were the opposite of a typical earnings profile, with a negative sign for the age relationship and a positive sign for the age-squared relationship. Therefore, the earnings function of the Newry Mill women was re-estimated for these years with the age-squared term omitted. As a result, their expected age-earnings profiles in Figure 2 and 3 are almost perfectly flat. Experienced female operatives had no market alternatives with which to force advancement relative to less experienced operatives, and so their age-earnings profile was as predicted from Figure 1(b). The family economy of the mature textile industry had created a surplus pool of female labor that could be allocated to positions in almost any fashion the mill desired.

The rapid build-up of demand prior to and just after America's entry into the war would have been expected to profoundly affect the depressed family economy of the Newry Mill. First, the surplus pool of female labor would eventually have been depleted. The supply of labor would have lost its infinite elasticity, as mills would have been forced to offer higher wages to attract additional labor, as well as hold on to their own. Secondly, experienced women operatives should now have been able to command higher earnings than inexperienced hands. With more job opportunities available, both in and out of the textile sector, both locally and in the larger cities, skilled hands could have taken higher-paying jobs elsewhere, or insisted on higher pay at home to keep them from leaving.

Furthermore, the supplemental tie of the women operative to her husband's mill job would have weakened. Should the husband have gone into the service or sought a position at the Norfolk, Virginia shipyards, the wife could choose to stay in the area `for the duration,' or follow him to be near his training camp or new job. Should the husband have remained at the mill, other local mills or businesses would be eager to hire her, with or without her spouse as part of the deal. As can be seen in Table 2, by Late 1942 only around 57 percent of the female employees sampled at Newry had their spouse working at the mill, compared to over 80 percent in Early 1937. The textile labor market for women should have begun to look more like the male market of Figure 1(a), and less like the family economy female market of Figure 1(b). As a result, a correlation between age and earnings should have appeared for the female labor force.

The age-earnings coefficients for the first and last quarters of 1942 can be found in Table 3. There was still no statistically significant relationship between age and earnings in Early 1942, but by this time, the normal positive coefficient with age and negative coefficient with age- squared had appeared. A statistically significant positive correlation in the age-earnings profile first appeared in the second quarter of 1942 (not shown in Table 3), and steadily developed into the typical earnings function of Late 1942. Figure 4 depicts the expected age-earnings profiles for the Newry female operatives for the first, second, and fourth quarters of 1942. By the end of 1942, women had projected earnings growth until approximately 40 years of age, followed by stagnation. The advancement opportunities for women were still limited relative to men, as can be seen in Figure 5 by the comparison of the female profiles of Figure 4 with the male earnings profile of the fourth quarter. Nonetheless, the family economy of Figures 1(a) and 1(b) had clearly been disrupted, and women were gaining a more independent role in the labor market of the Newry Mill.

Finally, Table 1 presents a close-up look at the job-advancement ladders for women at the Newry Mill during this period. A study of this changing job structure will highlight the sources of the growing correlation between age and earnings. First, the job ladder for Early 1941 shows the grouping of the majority of women's jobs at the minimum wage level of 33 cents an hour. The only job paying higher earnings was weaving, to which access was limited (65 percent male). Although the average age for weavers was greater than that for batteries and some miscellaneous positions, it was lower than the average age for spinners and spoolers. Moreover, the youngest operative in the sample was a weaver, and the average age of the weavers was essentially the same as the average of all the female employees (Table 2 �� mean of 32.06).

The first impact of the war can be seen in the Early 1942 job ladder by the increase of most wages above the new minimum wage level of 40 cents an hour. Only the battery position, a relatively unskilled job with a younger than average work force remained at the wage floor. This separation of the less-skilled battery earnings from those of the spinners and spoolers is evidence that the supply curve of more skilled labor had become separate and less elastic than for unskilled labor. Rising wartime demands then pulled skilled wages up faster than unskilled wages.

The second impact in Early 1942 is a large increase in the number of spinners at the mill, and their overall percentage of women's jobs. These positions were filled with younger women, lowering the average age of these workers from 35 to 32. Finally, although the number of women weavers did not increase substantially, their average age rose to the level of the other non-battery positions. There was still no statistically significant relationship between age and earnings (Table 3), but there is evidence that the mill was starting to sort younger women employees into the less-skilled spinning department, and older operatives into the weaving positions.

By Late 1942, rising demand had apparently hit the weaving department. There was a substantial increase in battery jobs, with an additional decline in the average age to 26. However, there was no corresponding increase in the number of female weavers, although a substantial increase in their average age, from 32 to 37. The mill was now definitely reacting to separate supply curves for different levels of experience in the labor market for women. Battery jobs were going to younger women on average, spinning jobs to slightly older women, and finally weaving positions to the highest average age level. A statistically significant relationship had been established for the age-earnings profile. Many women could now hope to advance up the job ladder with age. The family economy, at least for the duration, had been broken.


Of course, the most severe disruption of the Newry Mill economy was yet to come. It is important to note that at this early stage of the war, the number of male employees was also rising (Table 2). The correlation of women's earnings with age was not due to a necessary substitution into vacant male jobs. Women held roughly the same percentage of every job open to them in Late 1942 as they did in Early 1941. The difference was that the mill was now forced to move more experienced women into higher-paying positions or risk losing them. Nor was this correlation likely due to a large influx of young completely green hands having to be put into the simplest jobs. The cotton mill South had worked for over fifty years to create an entire class of people who had grown up in the villages and had at least some knowledge of mill work. The overall percentage of the locally-born operatives was essentially unchanged from Early 1937, and only slightly down from Early 1941 (Table 2). A large influx of truly new textile hands would have to wait until the Civil-Rights Era.

In conclusion, the Newry Mill management had not yet been physically forced in 1942 to abandon the family economy. If it had only been willing to pay high enough wages to its male operatives, it probably could have maintained this culture and its surplus female labor force, at least up to this stage of the conflict. It raised wage levels three times in 1942, but beyond that, the family economy was simply not worth enough to them to fight for it. Whether they believed the mill world would return to normal at war's end is impossible to say, but future research into jobs and earnings at the Newry Mill during and after the war will at least give some insight into what actually happened. Furthermore, analysis of the piece-rate productivity data of the Newry Mill can be used to study the issues of its sexually-based and age-based job assignments.

From hindsight, it seems difficult to believe that the Southern cotton mill family economy would really have been able to reestablish itself in the post-war South. It had always depended on a certain slackness of labor demand, either in textiles or the agriculture out of which it sprang. It had depended on a isolation that kept its family largely unaware of the opportunities that lay elsewhere, or at least unable or unwilling to grasp the significance of what lay beyond the village streets. In any case, the next three years "... really shook the place. ... Every young man must have gone. Newry was never the same."