Since the 1940s, the southern US has been transformed from a region of backward agriculture, low-wage industries located in small towns and rural areas, and unrelenting racial segregation into a modern society and economy. In 1950, there were no metropolitan areas in the South with a population of one million or more, but 18 had populations in excess of one million in 2000. The populations of the Atlanta, Dallas-Fort Worth, Houston, and Miami metropolitan areas grew to over 4 millions. Population growth in the 18 largest metropolitan areas accounts for 63% of the total population growth in the South from 1950 to 2000. The transformation of the South is, to a sizable extent, a transformation to an urbanized society. This paper documents this urbanization by examining population and employment growth in those 18 metropolitan areas.

1. Introduction

Beginning in the 1930s, the southern states of the US were transformed, according to the title of the book by Schulman [1], From Cotton Belt to Sun Belt. Wright’s [2] title is Old South, New South, and he states [2, page 241] that

By the 1980s (and indeed much earlier in many places), a new Southern economy prevailed, located in the same geographic space as the old one, but encompassing a very different package of labor, capital, natural resources, and entrepreneurship: not an advanced version of the old economy, but a new economy.

The transformation of the South was a goal of President Franklin D. Roosevelt. The Tennessee Valley Authority created in 1933 did not pursue aggressive economic transformation for its first five years but, as recounted by Schulman [1], did so beginning from 1938. The 1938 federal Report on Economic Conditions of the South stated that “The low income belt of the South is a belt of sickness, misery, and unnecessary death.” In that same year, FDR declared that the South was the nation’s number one economic problem. At that time the economy of the South was based on backward agriculture, low-wage industries located in smaller towns and rural areas, and unrelenting racial segregation and discrimination and denial of voting and other legal rights. This southern system was created in the decades after the Civil War in the interest of a ruling coalition—a result that Olson [3] argued is an application of his “logic of collective action.” All of this also was documented in Myrdal’s monumental work [4] An American Dilemma, a book that was featured prominently in Myrdal’s Nobel Prize citation. The year 1938 is also the date of passage of the Fair Labor Standards Act, which instituted a minimum wage of 30 cents per hour for the entire nation. This minimum wage had been set in 1933 under the National Industrial Recovery Act (NIRA), but the demise of this program had resulted in some wage reductions in southern industries. FLSA made the minimum wage permanent, and the minimum was raised to 40 cents per hour in 1940. Wright [2] argues that the New Deal policies in general (including the Tennessee Valley Authority) and the minimum wage in particular were important causes of the beginning of the transformation of the South. But it is clear that World War II was the watershed event. The war brought military bases and military production facilities to the South and drew workers from the farms and low-wage industries into this new economy. The end of the war left industrial workers and production facilities that could be used for defense contracts during the Cold War and for nondefense industry as well.

The dramatic changes in the economy of the South after World War II have been attributed to list of factors, as summarized by Roland [5, pages 11–20], Wright [2, page 239], and Glaeser and Tobio [6].(1)Federal spending, especially on defense and the space program, stimulated the southern economy—and continues to do so. Schulman [1] is one who emphasizes the role of the federal government and the influence of the southern congressional delegation. Schulman [1] points to the federal subsidization of transportation facilities—interstate highways and airports—which was particularly important for the South with its relatively poor transportation facilities.(2)The mechanization of cotton harvesting began in the late 1940s and was essentially complete by 1960. Labor was freed to work in nonagricultural jobs. Promotion of industry became a major purpose of state governments.(3)The growth of the region resulted from the movement of both capital and labor in response to economic incentives. This line of argument is the one that regional economists have explored at length, beginning with the work of Borts and Stein [7]. It turned out that the story of economic convergence at the state level for the 1929–53 period is complex and required more than a single sector model. For example, a study of southern industrial wages and employment changes by Fuchs and Perlman [8] showed that southern wages in the decade after World War II were relatively low because southern manufacturing was concentrated in low-wage industries (e.g., lumber and textiles) and because individual southern industries paid low wages. Furthermore, the South was tending to attract low-wage industries. The later movement of capital and higher-skilled labor to the South increased productivity, leading to further growth. Econometric work on structural transformation and regional convergence continues to this day, as exemplified by Caselli and Coleman [9] and Higgins et al., [10].(4)The South has benefited from its climate (after the installation of central air conditioning systems) and other amenities and has therefore attracted both highly skilled professional workers and retirees. Indeed, some of the major metropolitan areas in the South such as Tampa, Orlando, Fort Lauderdale, and West Palm Beach are classified as resort/retirement places by Stanback [11]. (Note that these are all in Florida.)(5)The South carefully cultivated a reputation for a good “business climate” consisting of an absence of labor unions and restrictive laws, aggressive marketing, and subsidies to industry that often took the form of favorable terms for land in industrial parks set up by local government. This point was made persuasively by Cobb [12] in The Selling of the South: The Southern Crusade for Industrial Development 1936–1990. Indeed, the southern techniques for “smokestack chasing” were exported to the rest of the nation—and achieved only limited success in other regions. The absence of labor unions and other impediments to business is sometimes called the “clean slate” argument and was applied to the South by Olson [3]. The use of industrial parks is widespread. The most prominent one is Research Triangle Park, established by Governor Hodges in 1959 in the Durham, Raleigh, Chapel Hill area of North Carolina. Another aspect of favorable business climate is, as Glaeser and Tobio [6, page 616] state it, a “high tolerance for new construction” of housing. Indeed, the econometric tests conducted by Glaeser and Tobio [6] lead them to conclude that elastic housing supply was an important factor in southern growth from 1980 to 2000.

Economic growth of urban areas can result from several forces: export demand, public and private capital investment, labor force growth, improvement in the quality of the labor force (education and training), technical change and innovation, entrepreneurship, agglomeration economies (one-time static effects and on-going dynamic effects), and amenities. Government policy can play a role in enhancing all of these sources of growth.

Wright [2] argues that, for all of these forces to operate, Southerners had to decide to change. They had to be made to realize that their old economy was no longer viable, and that their old way of life, eventually including the system of racial discrimination, had to change. The ruling coalition had to change the “system.” For Wright [2, page 270]:

The ironic conclusion to the story is that the only major act of conscious economic suppression by northern forces, the imposition of national wage and labor standards beginning in the 1930s, was the decisive step in abolishing the separate Southern economy. When southern property owners no longer had an economic stake in maintaining the separateness of the southern labor market, they opened the regional doors to much larger flows of outside labor and capital, with the result that the South as a distinct economic entity has all but disappeared.

What also have (almost) disappeared are the overtly racist, anti-urban Dixiecrat demagogues such as Governor Eugene Talmadge of Georgia, as quoted by Anderson [13], who declared in 1946 that public education “… ain’t never taught a man to plant cotton.” According to Anderson [13], Talmadge invited his rural male constituents: “Come see me at the mansion. We’ll sit on the front porch and p… over the rail on those city bastards.” Eugene’s son Herman also served as Governor of Georgia in the 1950s, and he was a rather different sort of fellow who supported economic modernization (but not integration of higher education in the state). A more recent holder of the office is, of course, Governor and President Jimmy Carter. However, a more recent essay by Wright [14] argues that the economic transformation of the South from 1940 up through the mid-1960s was pursued while maintaining racial segregation. It took the civil rights movement and federal laws to bring African American southerners into the economic and political mainstream. However, Wright’s [14, page 90] concluding statement is “But the biracial quality of southern life seems deeply entrenched and unlikely to disappear in the near future.”

One enters the topic of the transformation of the South with caution because so many scholars have devoted much time and effort to it. I come to the project as a Northerner and as an urban economist with some credentials in contemporary urban economic history, McDonald [15]. The purpose of this paper is to examine the evolution of the major urban areas of the South during 1950–2000 in order to understand the roles that these metropolitan areas played in the economic transformation. The idea is to bring a fresh perspective to the study of transformation of the South by concentrating on the southern metropolitan areas as, to use Roland’s [5] term, the “improbable era” unfolded. One historian of the urban South, Goldfeld [16], suggests that urban areas reflect the history and traditions of their regions and that, since the South is the most distinct region in the US, it is inevitable that southern urban areas are different from their northern and western counterparts. Goldfeld [16, page 1033] states that the South is a distinct region in its history of “… ruralism, race, and colonialism,” so that its modern cities also will have distinct characteristics. Can the urban economist find such distinctions in the data? In contrast to Goldfeld [16], McKinney and Bourque [17] documented that the South after World War II up to 1980 was becoming more like the rest of the nation in a wide variety of economic measures—urbanization, industrialization, occupations, income, and education. What about the subsequent thirty years?

2. Urbanization of the South: Some Basic Data

This section presents the basic urbanization story. Table 1 shows the population of the South by state for 1950 to 2000 (in thousands). This table shows that there are sharp distinctions among the states, and that the fifty years divide into two periods. During 1950 to 1970 the population of the South grew only slightly more rapidly than did the population of the nation, largely because these were the final decades of the “Great Migration” of African Americans from the South to the North. The total population of the South increased by 36.9% (from 36.55 million to 50.05 million) as the nation’s population grew by 34.1%. The African American population of the South increased only 12.5% from 9.05 million to 10.18 million, while the African American population of the nation increased by 50.0% during the same twenty years. The percentage of the African American population that lived in the South fell from 60.2% in 1950 to 45.1% in 1970. See Tolnay [18] for a survey of research on the Great Migration—and the return migration from that began in the 1970s.

The data for the states show that 55.6% of the population growth from 1950 to 1970 in the South was attributable to two states—Florida and Texas. Virginia, North Carolina, Georgia, and Louisiana grew by 958,000 to 1.33 million people, but the five states of South Carolina, Tennessee, Alabama, Mississippi, and Arkansas had population growth of 632,000 or less. Four of those states—South Carolina, Alabama, Mississippi, and Arkansas—had fewer African American residents in 1970 than in 1950. As it happens, South Carolina, Mississippi, and Arkansas are the three states in the South that did not have a metropolitan area of at least one million in population in 2000 (and the only such metropolitan area in Alabama, Birmingham, is the old industrial city in the South that grew relatively slowly).

Population growth in the South greatly exceeded the nation’s growth after 1970. The South increased by 68.4% (from 50.05 million to 84.28 million) as the nation increased by 38.3%; the share of the population of the nation that lived in the South increased from 24.8% to 30.1%. The African American population of the South increased to 61.0% as the number of African Americans in the nation grew by 53.4%. The share of African Americans residing in the South fell from 60.2% to 45.1% during 1950–1970, but that share increased to 47.4% in 2000. Once again Florida and Texas accounted for a majority of the population growth in the South: 55.0% to be exact, which is virtually identical to the 55.6% figure for 1950–1970. Both of these states added more than 9 million people from 1970 to 2000, but the nature of that population growth differed. In 2000, 17.6% of the population of Florida was with age 65 and over, compared to 12.4% for the nation and the entire South (and no other state in the South comes close to Florida’s figure for percentage elderly). This means that Florida was home to 830,000 more elderly people than it would have had if its percentage matched the southern and national average. Florida is, of course, a primary destination for retirees from around the nation. In contrast, only 9.9% of the population of Texas in 2000 was age 65 or older. Florida grew from 2.77 million in 1950 to 15.98 million in 2000, which is growth of 477%. Texas grew by 170% over this same period, compared to 85.5% for the nation. Three other states grew more rapidly than the nation; Georgia added 138% to its population, Virginia increased by 113%, and North Carolina grew by 98%. The population of South Carolina grew by 89.5%, which is only slightly faster than the national average. The other five states of the South had population growth that was appreciably slower than the nation’s growth. These states include Tennessee, Alabama, Mississippi, Louisiana, and Arkansas, and population growth ranged from 72.8% in Tennessee to 30.6% in Mississippi. The data for the states are illuminating, but they do not show clearly the influence of metropolitan areas.

The 18 metropolitan areas in the South with a population of at least one million as of 2000 are shown in Figure 1. The charts on the map display the population of the metropolitan area in 1950, 1970, and 2000. The charts are scaled in units of one million. For example, the population of the Dallas-Fort Worth metro area in 1950 is slightly less than one million, over two million in 1970, and over five million in 2000. The population of Memphis was about one-half million in 1950 and over one million in 2000. The 18 metropolitan areas are as follows: Norfolk, Portsmouth, Newport News, VARichmond, Petersburg, VACharlotte, Gastonia, Rock Hill, NCGreensboro, Winston Salem, High Point, NCRaleigh-Durham (with Chapel Hill), NCAtlanta, GAJacksonville, FL Miami, Ft. Lauderdale, West Palm Beach, FLOrlando, FLTampa, St. Petersburg, FLMemphis, TNNashville, TNBirmingham, ALNew Orleans, LAAustin, San Marcos, TXDallas-Ft. Worth, TXHouston, TX (without Galveston)San Antonio, TX.This list of metro areas is organized as follows. On the map begin with Virginia (Norfolk and Richmond); proceed down the coastline to North Carolina (Charlotte, Greensboro, Raleigh-Durham), Georgia (Atlanta), and Florida (Jacksonville, Miami, Orlando, Tampa); go back to the North to Tennessee (Memphis and Nashville); go South to Alabama (Birmingham) and Louisiana (New Orleans); lastly go West to Texas (Austin, Dallas, Houston, San Antonio). The geographic definitions of the metropolitan areas determined by the US Bureau of the Census (USBC) are a group of counties, with at least one designated as the central county. In this paper, each metropolitan area is defined as it was by the USBC at the time of the census. In some cases for 1950, metropolitan areas were combined to match the later definition. For example, Durham and Raleigh were separate metropolitan areas in 1950, but were one metro area in the later definitions. Population data for each of the metro areas from 1950 to 2000 are displayed in the Table 5.

Over the 50 years from 1950 to 2000, the population of the South increased by 130.6% and the 18 metropolitan areas had population growth of 361% (from 8.33 million to 38.37 million). These 18 metropolitan areas account for 63% of the population growth in the South from 1950 to 2000 and 59% of the growth from 1970 to 2000. The fraction of the southern population that lived in those metropolitan areas doubled from 22.8% to 45.5% even as the population of the region was increasing rapidly. The South did not contain a metropolitan area with a population of one million or more in 1950, but 18 had populations in excess of one million in 2000. The North contained 12 metropolitan areas with population in excess of one million in 1950, and the West had two (Los Angeles and San Francisco-Oakland) such areas in 1950. The 18 metropolitan areas had a total population of 8.33 million in 1950, which was 22.8% of the population of the South at that time. The largest metropolitan area was Dallas-Fort Worth, with a population of 976,000, and Houston was second with 808,000 people. Orlando, Florida had only 115,000 people, and Austin, Texas was home to just 161,000. The Miami area had 694,000 people, and metropolitan Atlanta boasted just 672,000.

The population of the 18 metropolitan areas more than doubled from 1950 to 1970 to 18.24 million as the population of the South increased by 36.9%. In 1970, 36.5% of southerners lived in these 18 metropolitan areas. Seven of the metropolitan areas reached the million-person mark, and three—Miami, Dallas-Ft. Worth, and Houston—reached two million. These three metropolitan areas accounted for 30.7% of the population growth in the 18 metro areas (4.14 million out of 13.49 million), but all grew by at least 58.0% (Birmingham). Orlando almost tripled in size from 115,000 to 428,000. The South was becoming urban.

The African American population of the South was becoming urbanized as well. In 1950, 22.3% of the African American population of the South lived in one of the 18 metropolitan areas, compared to 22.8% for the entire southern population. By 1970 the African American population in the 18 metropolitan areas had grown from 1.84 million in 1950 to 3.42 million, and the percentage of the southern African American population in the 18 metro areas had increased to 33.6% (compared to 39.6% for the total southern population). The African American population of the South was becoming “urban” at a less rapid rate than the southern population as a whole. The largest metropolitan areas—Miami, Dallas-Ft. Worth, and Houston—all had increases in the African American population in excess of 200,000, and together accounted for 40.7% of the increase in the African American population in the 18. All of the 18 metro areas experienced increases in the African American population, although the increases were quite small in Birmingham and Austin.

During the next thirty years the population in the 18 metro areas continued to grow very rapidly—from 18.24 million to 38.37 million (an increase of 103.6%, compared to overall southern population growth of 68.4%). In contrast, McDonald (15) shows that the population of the Northeast grew by just 12.5%, and the average growth for the top 17 metropolitan areas in the Northeast was 14.8%. The population of the West grew by 82% during these decades, and the top six western metro areas grew by 94%. In short, the population of the South was concentrating in its top metropolitan areas at a rate that exceeded those in the other regions of the nation. In 2000, the population of Miami and Dallas-Ft. Worth both exceeded five million and Houston and Atlanta exceeded four million. Population growth in each of the top four exceeded 2.4 million over the thirty years, and together their growth accounted for 53% of the growth in the total population of the 18 metro areas and 31% of the growth in the South. However, all of the 18 experienced growth; Tampa emerged in fifth place in the South with 2.4 million people, Orlando boomed from 453,000 in 1970 to 1.64 million in 2000, and Austin shot up from 323,000 to 1.25 million. Eleven metro areas joined the million-person club. But population grew by only 37% in Birmingham. In 2000, 45.5% of all southerners lived in one of the 18 metro areas.

Growth of the African American population in the 18 metro areas of 123% exceeded the population growth of the 18 metro areas of 103.8%. The percentage of the population in the 18 metro areas that were African American had declined from 22.0% in 1950 to 17.3% in 1970, but increased to 19.7% in 2000. The percentage of African American southerners who lived in one of the 18 metro areas increased from 33.6% in 1970 to 46.5% in 2000, a figure which slightly exceeds the percentage of all southerners (45.9%) who lived in the 18 metro areas. For the first time the African American population of the South was, by this measure, more highly urbanized than was the entire population of the South.

The growth of the Hispanic population was also an important factor in the 1990s. The percentage of the population of the 18 metro areas that was Hispanic in 1990 stood at 11.8%. The Hispanic population was concentrated in Miami, Dallas-Ft. Worth, Houston, and San Antonio. During the 1990s, the Hispanic population in the 18 metro areas increased from 3.68 million to 6.51 million (76.9%), which increased the proportion of the Hispanic population in these metro areas to 16.8%. Growth in the Hispanic population was 37.3% of the population growth in the 18 metro areas in the decade. All 18 experienced growth in this population group, but most of the growth (59.4%) took place in Miami (584,000), Dallas-Ft. Worth (507,000), and Houston (590,000). The urban South is becoming increasingly Hispanic, and this trend is primarily located in, but not confined to, Texas and Florida.

3. Metropolitan Growth Rates

Regression models were estimated to determine the variables that were associated with metropolitan growth. Econometric studies have examined population, employment, and income growth rates for all metropolitan areas in the US, and a survey of these studies appears in McDonald and McMillen [19, pages 448-449]. A measure of educational attainment of the adult population is found to be a positive and statistically significant factor in every study. The measure of educational attainment that is used most frequently is the percentage of adults who are college graduates. No other variable produces such consistent results. All studies include measures of the composition of employment in the base year such as percentage of employment in manufacturing—a variable that is found to have a negative effect in most studies, but not in the recent studies by Green [20] and Blumenthal et al. [21]. Also, all studies include a measure of the size of the metropolitan area in the base year, but this variable enters with a positive sign in some studies and a negative sign in other studies. Some studies include measures of weather (average January and/or July temperature, heating degree days, etc.), and warmer temperatures are associated with more rapid population growth. (These variables are not included here because all of the metro areas are in the South.) The racial composition of the population was included only by Blumenthal et al. [21], with the result that the percentage of the population that is African American in the base year is negatively associated with employment growth.

The dependent variables are percentage population growth, percentage employment growth, and percentage growth in inflation-adjusted median family income growth from 1970 to 2000 as a decimal such as 0.75 for 75% (PctPopGrow, PctEmpGrow, and PctIncGrow), and the right-hand-side variables are population in 1970 (Pop70) in millions, percentage of adults with college degrees (PctGrad), and percentage of the population that is African American (PctAfAm). Summary statistics for these variables are shown in Table 2. Mean percentage population growth was 118%, mean percentage employment growth was 137%, and mean percentage real income growth was 27.5%.

The estimated equation for population growth is The -squared for the estimated equation is 0.700 (adjusted -squared 0.636), and unsigned values are in parentheses. The critical value for the statistics (95% level) with a sample size of 18 is 2.12. The results show that the size of the metro area in 1970 was a statistically significant variable ( statistic of 2.31), with a negative sign. Smaller metro areas in 1970 such as Raleigh-Durham, Orlando, and Austin that exceeded one million in population in 2000 grew more rapidly in these thirty years. Population growth was statistically significantly ( statistic of 4.06) and positively related to the percentage of adults with college degrees in 1970, a result that is consistent with previous studies of metropolitan population growth. The effect is quite large; an increase in percentage college graduates of 1.0 percent in 1970 was associated with an increase in population growth over thirty years of 22%. The percentage of the population that is African American is negatively associated with population growth, and this effect is statistically significant as well ( statistic of 2.41). An increase in percentage African American in 1970 was associated with a reduction in population growth of 3.5%. The simple correlation between the percentage college graduates and the percentage African American in 1970 is . This is not a particularly high level of collinearity, and so the regression model is able to distinguish between these two variables.

The finding regarding college graduates is as expected, but the results for African American population are somewhat surprising—but consistent with Blumenthal et al. [21]. The three metropolitan areas with the highest percentage of African American population in 1970 are Birmingham (29.4%), Memphis (37.5%), and New Orleans (31%). These metropolitan areas had relatively low population growth figures for the 30 years of 37%, 44%, and 26%, respectively. The underlying reasons for this finding will require further research.

The estimated equation for percentage employment growth is The -squared for the estimated equation is 0.635 (adjusted -squared of 0.5575), and unsigned statistics are in parentheses. The results are very similar to the equation estimated for population growth; employment growth is positively related to college graduates, negatively related to African American population, and negatively (but not statistically significantly) related to the size of the metropolitan area. An increase in the percentage of adults with college degrees in 1970 was related to an increase in employment growth of 18.7% over thirty years from 1970 to 2000, and an increase in African American population of one percentage point was associated with lower employment growth by 4.8% from 1970 to 2000.

The estimated equation for inflation-adjusted median family income growth is The -squared for the estimated equation is 0.653 (adjusted -squared of 0.607), and unsigned statistics are in parentheses. Median income growth is positively related to college graduates and negatively related to the size of the metropolitan area. These results are similar to the findings in the study of median real wage growth by Glaeser and Saiz [22]. Median income growth is unrelated to percentage African American population, and so this variable was dropped from the estimated equation.

Alternative specifications of the three models were estimated; population growth, employment growth, and median family income growth in the southern metropolitan areas were unrelated to the composition of employment (percentage of employment in manufacturing; public administration; finance, insurance, and real estate and professional services), median income in 1970, and the percentage of the population in poverty in 1970. The finding that an adult population with more college graduates is related to growth (in population, employment, and real family income) is consistent with the existence of several sources of growth: quality of the labor force, technical change and innovation, and entrepreneurship.

4. The New Urban Economy of the South

How has the nature of the southern economy changed? Some researchers have examined the nation’s urban areas and reached some conclusions about the 18 metropolitan areas in this study. Bartley [23], in his comprehensive history of the new South, offers a brief characterization of the main metropolitan areas in 1980. Stanback [11] classified metropolitan areas by type and size for 1990 based on location quotients for employment. Florida [24] examined the occupational patterns in metropolitan areas as of 1999 and ranked them according to his measure of the “super-creative class,” which consists in the following occupations:Computer and mathematical occupationsArchitecture and engineering occupationsLife, physical, and social science occupationsEducation, training, and library occupationsArts, design, entertainment, sports, and media occupations.

Florida [24] computes the percentage of total employment engaged in these occupations.

A summary of their findings for southern metropolitan areas is displayed in Table 6.

Bartley [23] noted that, in 1980, Atlanta and Dallas-Ft. Worth had become the commercial capitals of their regions: the Southeast and the Southwest. In his view, they sit at the top of the regional urban hierarchy. Seven other metropolitan areas are considered to be regional commercial capitals; these include Richmond, Charlotte, Jacksonville, Memphis, Nashville, Birmingham, and New Orleans. Three are resort/retirement areas—Miami, Orlando, and Tampa, although the economy of Miami is becoming increasing diverse. Two, Norfolk and San Antonio, were still considered to be dominated by the military. Raleigh-Durham is dominated by the Research Triangle and the nearby universities, Houston is the nation’s center of the petrochemical industry, and Austin was considered (erroneously) just to be a state capital.

Nearly all of Stanback’s [11] data-driven classifications for 1990 are consistent with Bartley’s [23] informal assessments. Atlanta and Dallas-Ft. Worth are large diversified service centers. Houston is rated as a large diversified service center as well, and so Stanback [11] finds that Houston is a more diversified economy than Bartley [23] credits. Bartley’s regional capitals are Stanback’s diversified service centers. Norfolk and San Antonio are government and military centers, and Miami, Orlando, and Tampa are classified as resort/retirement centers (with Miami considered to be a diversified service center, as well). Raleigh-Durham is classified as a government/service center, and Greensboro (not mentioned by Bartley) is a production center.

Florida’s [24] enumeration of people employed in super-creative occupations in 1999 gives a different perspective on the 18 metropolitan areas. Raleigh-Durham is number one among the top 50 metropolitan areas in the nation with 16.9% of its workers in the super-creative class, and Austin is number five. The others in the top half of the top 50 are Houston (rank of 11), Atlanta (15th), Norfolk (18th), and Dallas-Ft. Worth (22nd). However, the other 12 metropolitan areas in the South are in the bottom half of the largest 50 metropolitan areas, and six of those 12 rank in the bottom ten of the 50 largest metro areas. Those with these low ranks are (starting at the bottom) Memphis, Nashville, Miami, New Orleans, and Charlotte and Birmingham (tied). According to Florida [24], a low ranking on super-creative people does not bode well for future economic vitality. Given the rapid growth that southern metropolitan areas have experienced, one can certainly question the validity of this hypothesis for the South.

Data showing the evolution of employment from 1950 to 2000 by major industry category for the 18 metropolitan areas are in Table 7. Total employment rates in 1950, 1970, and 2000 are shown, along with employment percentages for manufacturing; an aggregation of FIRE (finance, insurance, and real estate), business and repair services, and professional services (including health care and public and private education); public administration. Employment was reported consistently by the Census Bureau using the old SIC code system.

The employment data show that total non-agricultural employment in the 18 metro areas increased from 3.18 million in 1950 to 7.53 million in 1970 and then to 17.66 million in 2000. Along with this huge growth in total employment came huge growth in manufacturing employment as well. Total manufacturing jobs increased from 600,000 in 1950 to 1.53 million in 1970 (increasing its share of total employment in the 18 metro areas from 18.9% to 20.4%). During this same time period manufacturing employment in US increased from 14.0 million to 17.8 million, but declined as a percentage of total employment from 30.9% to 25.1%. Manufacturing employment in the southern metro areas constituted 24% of the total increase in manufacturing in the nation. While a good deal of manufacturing employment serves the local and regional market, this shows the success of the “selling of the South” to manufacturers during the period. Also, please remember that this growth in manufacturing took place in the larger metropolitan areas, not in the rural areas and small towns. All 18 metro areas added manufacturing jobs during these twenty years, with the largest increases (at least 96,000 jobs) in the usual suspects of Atlanta, Miami, Dallas-Ft. Worth, and Houston. The manufacturing sector added another 682,000 jobs in the 18 metro areas from 1970 to 2000, but this increase of 44.5% compared to an increase in total employment of 134.7% means that the fraction of jobs in manufacturing fell from 20.4% to 12.5% (compared to 25.1% and 13.1% for the nation.) The southern metro areas were never as highly concentrated in manufacturing employment as the nation, but by 2000 these two percentages were almost identical.

Manufacturing jobs declined in Birmingham (the old industrial metro area) and New Orleans and increased very little or not at all in Richmond, Greensboro (another older production center), and Memphis. Manufacturing job growth was concentrated in Atlanta, Miami, Austin, Dallas-Ft. Worth, and Houston. Manufacturing jobs in Austin increased from 159,000 to 662,000. Overall manufacturing declined as a share of employment after 1970 because service sector jobs grew even more rapidly.

Employment growth in the metropolitan areas of the South was dominated by growth in the newer services sectors of FIRE, business and repair services, and professional services (including education). These sectors made up only 13.3% of total employment in the 18 metro areas in 1950 (compared to 19% for the nation), but their proportion increased to 24.6% in 1970 and 38.7% in 2000. These figures for 1970 and 2000 are close to the national percentages of 25.9% and 37.1%. The urban South joined the modern service economy with an amazing increase in employment of 6.41 million jobs over 50 years. Professional services (including health services, education, and others) led the way with an increase from 186,000 jobs to 3.89 million jobs, a percentage increase that is ridiculously too large to report.

The final category reported in the appendix is public administration. In 1950, public administration employment was, as one would have expected, an especially large fraction of employment in Norfolk (17.6%), Austin (small state capital at the time, with 18.8%), and San Antonio (15.8%). Employment in this category was 7.9% of total non-agricultural employment in the 18 metro areas. Public administration employment increased in the 18 metro areas by 58% from 1950 to 1970 and 98% from 1970 to 2000, but declined as a fraction of total employment from 7.9% in 1950 to 5.3% in 1970 and 4.4% in 2000. The South was relying less and less on public administration jobs as time passed.

What more general conclusions do these data suggest? The economies of the metropolitan areas of the South grew rapidly from 1950 to 1970, and the growth in manufacturing employment matched overall employment growth during this period. However, in the subsequent 30 years, the growth of the service economy took over as the main driver of employment growth. Manufacturing employment continued to grow to the end of the century, but the metropolitan areas of the South clearly were becoming part of the new service economy—and relying to a lesser extent on public administration employment. The four mega-metro areas of Atlanta, Miami, Dallas-Ft. Worth, and Houston differed somewhat in their industry employment shares. Atlanta was very close to the means for the 18 metro areas with 12.7% in manufacturing, 36.6% in the new service economy, and 4.3% in public administration employment. The Houston economy presented a very similar picture. Miami had a smaller fraction of jobs in manufacturing (7.8%) and more jobs in the new service economy (39.4%), while Dallas-Ft. Worth had more in manufacturing (14.5%), the same fraction in the new service economy (36.7%), and less (2.9%) in public administration.

5. Social Indicators in the Southern Metropolitan Areas

How have the people of the 18 metro areas fared during this time of profound transformation and rapid growth? Comparisons can be made on a variety of indicators for 1970 and later. A preliminary look at this question is provided in this section, and detailed data are in Table 8, which displays several social indicators for 1970 and 2000.

Median family income (in 2005 dollars) increased substantially in all 18 metro areas. The mean increase for the 18 metro areas was 27.1% over thirty years. The largest increase (56.1% in 30 years) was in Austin, which boomed as a high-tech center in the Southwest. The smallest increases were in the Miami area (9.9%), which experienced a heavy influx of immigrants from abroad, and in the old metro area of New Orleans (8.4%). Indeed, as McDonald [25] describes, economic and social conditions were not progressing well in New Orleans before Katrina struck in 2005. As noted earlier, growth in median family income was associated positively with the percentage of adults with college degrees in 1970. The mean poverty rate for the 18 fell from 16.5% in 1970 to 12.3% in 2000, although poverty was greater in 2000 than in 1970 in the Miami area and in Houston. Once again, the influx of immigrants may have been the cause for this outcome in these two metropolitan areas.

Another measure of poverty is the number of people living in areas of concentrated poverty, defined by Jargowsky [26] as census tracts with 40% or more of the population in poverty. The findings for concentrated poverty from Jargowsky [26, 27] and tell a mixed story. The number of people living in areas of concentrated poverty increased in Atlanta, the Miami area, Austin, and Houston. Sizable declines in concentrated poverty occurred in the Norfolk area, Charlotte, Tampa, Memphis, Birmingham, and San Antonio.

Another widely used indicator of possible social problems is the percentage of families with children with only one parent present. This indicator has increased dramatically in the nation as a whole, and the urban South is no exception. The mean value for this indicator in the 18 metro areas more than doubled from 14.2% in 1970 to 30.1% in 2000, and the story is pretty much the same in all of the 18.

Table 8 provides data on the percentages of African American residents of each of the 18 metro areas. A new piece of information is the increase in the foreign-born population from a mean of 2.8% of the population in 1970 to 9.6% in 2000 (not shown). The foreign-born population is concentrated in the Miami area with 35.0% of the population in 2000 (up from 17.2% in Miami in 1970) but exists in significant numbers in most of the other 17 metro areas as well.

As one would expect, rapid population growth means low housing vacancy rates in general. The mean vacancy rate for the 18 was 5.8% in 2000 compared to the national figure of 8.0%. The highest vacancy rates in 2000 were in Birmingham (8.1%) and New Orleans (8.0%), and the lowest was in Austin (3.9%). A vacancy rate of 8% in a metropolitan area can mean a high vacancy rate in the central city, and indeed this is the case. The vacancy rate in 2000 was 11.5% in both the Birmingham and New Orleans central cities. Housing construction in the suburbs was, through the filtering effect, producing vacancies in the central city. This is a phenomenon more often seen in the major urban areas of the North.

The last entry in Table 8 is the percentage of adults with college degrees. The mean value for the 18 metro areas in 1970 is 10.8%, and the mean for 2000 is 26.2%. This increase in college education is in line with the national trend; 11% in 1970 and 25.6% in 2000. It is no surprise that Austin had the greatest percentage of college degrees in 1970 (16.3%) and ranked second to Raleigh-Durham with 36.7% in 2000. Birmingham (8.3%), Jacksonville (8.8%), and Memphis (8.9%) had lowest college graduate percentages in 1970. The three metro areas with the lowest percentages in 2000 are Tampa (21.7%), Memphis (22.0%), and San Antonio (22.1%). Recall that growth in population, employment, and family income from 1970 to 2000 were strongly related to the percentage of college graduates in 1970.

The basic finding is that the major metropolitan areas of the South did well in the last three decades of the 20th century, but the fruits of that success did not extend to all. Median incomes were up and poverty rates generally were down, but a few central cities struggled with increasing concentrations of poverty. Single-parent families doubled everywhere. Housing markets generally were tight, except in the central cities of Birmingham and New Orleans. The increase in the percentage of adults with college degrees is large and follows the national trend, but the southern metro areas in 2000 varied widely in this predictor of future growth—from a low of 21.7% in Tampa to 37.6% in Raleigh-Durham.

6. Changes in the US Urban Hierarchy

The urbanization of the South and urban growth in the West changed the hierarchy of urban areas in the US period Table 3 displays the top 25 metropolitan areas in 1950 and 2000.

The southern metro areas are shown in bold type. Los Angeles replaced Chicago as the second-largest metro area and San Francisco-Oakland moved up from number 7 to number 4. Membership in the top ten changed. Detroit, Pittsburgh, Cleveland, and St. Louis dropped out of the top ten and were replaced by Dallas-Fort Worth, Miami, Washington, DC, and Houston. Atlanta moved up from number 25 to number 12. In short, the southern metro area with the highest rank in 1950 was Dallas-Fort Worth at number 15, and in 2000 the South had four metro areas in the top 12. In addition, Tampa entered the top 25 at number 20, but New Orleans dropped out of the top 25. In addition the western metro areas of Phoenix and Denver entered the top 25 and Seattle moved up in the rankings from 21 to 14.

The evolution of urban hierarchies is a topic of a great deal of recent research. For example, see Bretagnolle and Pumain [28] for simulation models of the urban networks of Europe and the US and discussion of the recent literature. Their maps for 1940 and 2000 clearly show the emergence of Atlanta, Miami, Dallas-Fort Worth, and Houston as major metro areas. However, their simulation model indicates that New Orleans would have been the largest metro area in the South in 2000. Gaining an understanding of how urban networks evolve is a work in progress. Perhaps this paper will provide some data for these efforts, but formal urban hierarchy research is beyond its scope, which is to review the history of southern urbanization.

7. Conclusion

One approach to a conclusion is to compare the 18 largest metro areas of the South to the 16 largest metro areas in the North (excluding New York), which can be done using data from Wright [14]. New York is excluded on the grounds that it is much larger than any of the other metro areas, and it has functions in the national and world economies that no other metropolitan area performs. A few comparisons for 1970 to 2000 are contained in Table 4.

The table shows that population growth and employment growth were much slower in the northern metro areas, although employment growth exceeded population growth by 25.2% because females and baby boomers were entering the workforce in large numbers. This effect was less pronounced in the southern metro areas. Median family income in real terms advanced more rapidly in the southern metro areas as one would expect, and the poverty rate declined. The poverty rate in the northern metro areas was essentially unchanged comparing 1970 to 2000 (but it had been higher during the urban crisis period of 1970 to the early 1990s). The mean poverty rate in the southern metropolitan areas was 16.5% in 1970 compared to the average for the northern metro areas of 9.3%. The difference in the average poverty rates closed in the next thirty years, but the average southern poverty rate of 12.3% was still greater than the average poverty rate in the northern metro areas of 9.7%. The metro areas in the two regions seem to be converging on this measure. Lastly, we see that manufacturing as a fraction of employment fell by 7.8 percentage points in the southern metro areas and by 11.8% in the northern metro areas. The metropolitan areas of the North were still more heavily concentrated in manufacturing than were the metro areas of the South in 2000 (15.0% versus 12.3%), but the metropolitan areas in the two regions appear to be converging on this measure as well.

In short, while the southern metropolitan areas grew at a much more rapid pace than their northern counterparts, on some measures (median family income, poverty, and the industry composition of employment) the two groups of metro areas are converging. Median family income increased more rapidly in the southern metropolitan areas, further closing the gap with the northern metro areas. The poverty rate in the southern metro areas is converging to the northern poverty rate, while the fraction of employment in manufacturing in the northern metro areas seems to be converging to the smaller fraction observed in southern metro areas.

The next decade—2000 to 2010—is not discussed in this paper because this decade surely requires its own separate study. The decade was marked by the housing market bubble and crash, the financial crisis, and the deepest recession since the 1930s, all phenomena that were avoided in the previous five decades.


The author thanks Gavin Wright, the referees, and the editor for helpful comments.