Content uploaded by Byron Shafer
Author content
All content in this area was uploaded by Byron Shafer
Content may be subject to copyright.
B.J.Pol.S. 31, 601–625 Copyright 2001 Cambridge University Press
Printed in the United Kingdom
The Transformation of Southern Politics
Revisited: The House of Representatives as a
Window
BYRON E. SHAFER AND RICHARD G. C. JOHNSTON*
The appearance of a Republican majority among Southern members of the US House of
Representatives provides a substantive reason to reconsider political change in the post-war
South. And a newly created, merged dataset makes it possible to address the central and recurrent
propositions about change in Southern politics in a manner not previously possible. When this
is done, four basic contributions are highlighted, each a clear modification of the standard story.
The main impetus for partisan change proves to be economic development, and a changing
politics of economic interest. The main brake on this impetus proves to be legal desegregation,
and a changing politics of racial identity. Several indirect interactions of race and class then
enhance the impact of both contributions, while their joint impact is also powerfully shaped by
the strategic response of partisan elites: by the appearance of Republican challengers but,
especially, by the practical resistance of Democratic incumbents.
High on the list of transformative changes in American politics during the
post-war years is partisan change in the American South. Given the historical
background – a populous one-party region, unshakeable in attachment (and
behaviour) for nearly a century beforehand – such change would loom large
merely on its own. Yet serious partisan change in the American South inevitably
meant a change in partisan balance for the nation as a whole. Such change
implied a fundamental shift in the operational character of both political parties.
It implied a shift in the social coalitions associated with them. It necessitated
(two) shifting party programmes. It meant that all the major institutions of
American national government would work differently in its aftermath.
1
* Nuffield College, Oxford, and Department of Political Science, University of British Columbia,
respectively. Daniel L. Singer and Stuart N. Soroka were absolutely central to the creation (and
manipulation) of the dataset upon which this article is based. They have constructed it with a view
towards placing it in the ICPSR data archive. Charles O. Jones gave us a very insightful early reading,
which enriched the original paper enormously, while Geoffrey Evans and Adam Sheingate provided
a late check. The Departments of Political Science at the University of Virginia, the University of
Wisconsin and Vanderbilt University, and especially the Senior Research Seminar at Oxford
University, then provided opportunities to test its argument. We thank them all. An appendix on the
data analysis appears at the end of the web version of this article.
1Earl Black and Merle Black, Politics and Society in the South (Cambridge, Mass.: Harvard
University Press, 1987); Numan V. Bartley, The New South,1945–1980 (Baton Rouge: Louisiana
State University Press, 1995); Nelson W. Polsby, ‘A Revolution in Congress?’ (Inaugural lecture,
Oxford University, December 1997); Richard F. Fenno Jr, ‘Home Style Revisited: A Narrative Case
602 SHAFER AND JOHNSTON
The coming of serious partisan change to the South had been predicted –
dreamed about, really – since at least the days of William Jennings Bryan. It had
seemed closer in the interim, momentarily, each time the ‘solid South’ showed
some fresh crack in its presidential voting. But the end of that solidarity was not
inescapably at hand until Republicans began to make inroads into congressional
representation, and perhaps not until they actually captured a majority of
Southern seats in the House of Representatives. Fortunately, if it was partisan
change in the House that marked the undeniable arrival of national Republican-
ism, that body has additional major advantages as a window on the underlying
causes of this change. There are sufficient House districts to allow an analysis
that is not idiosyncratic to personalities. And there are sufficient polling data –
over time and across districts – to allow temporally extended, merged datasets,
joining individual and collective characteristics.
Accordingly, it is possible to work with composite district demographics and
specific personal backgrounds, elite candidate provision and mass public
response, aggregate congressional outcomes and individual voting behaviour.
This article sets out to do exactly that, tracing the rise of Southern Republicanism
in the House of Representatives and searching for the main causal contributions
to it. Where it differs from much earlier work is in having a concrete institutional
outcome as its focus, in wins and losses for the House; in organizing this analysis
around the two great changes in Southern social structure across the post-war
years, namely legal desegregation and economic development; in highlighting
the place of critical intermediaries – political parties and partisan elites – that
converted structural change into electoral victory (or defeat); and in using the
best available data for addressing all of the above.
Three principal datasets were merged to permit this analysis, and an attempt
to take these data seriously is critical to what follows:
(1) The cumulative file for the American National Elections Studies, mostly
tapped for individual-level data. This was supplemented with the individual
files for 1952, 1956, 1958 and 1960.
2
(2) The decennial census, mostly used for aggregate demographics of
congressional districts. In earlier years, these were drawn from the
Congressional District Data Book.Inlater years, they were drawn instead
from the decennial volumes on congressional districting by Congressional
Quarterly.
3
(3) A specialised compilation from the Inter University Consortium on Political
(F’note continued)
Study of Representational Change’ (paper presented at the Annual Meeting of the American Political
Science Association, Boston, Mass., 1998).
2National Election Studies, 1948–98 Cumulative Data File (Ann Arbor: University of Michigan,
Center for Political Studies, 1999); National Election Studies, 1952, 1956, 1958 and 1960 (Ann
Arbor: University of Michigan, Center for Political Studies, 1953, 1957, 1959 and 1961).
3Congressional District Data Book (Washington, DC: U S Department of Commerce, 1961,
1973, and supplements); Congressional Districts in the 1970s, 1980s and 1990s (Washington, DC:
Congressional Quarterly Inc., 1973, 1983 and 1993).
The Transformation of Southern Politics, Revisited 603
and Social Research, entitled Elections to the United States House of
Representatives, 1898–1992. For earlier years, this was supplemented with
the Guide to US Elections; for later years, with the compilations by
Congressional Quarterly.
4
ALTERNATIVE ENGINES FOR POLITICAL CHANGE
Despite seventy years of partisan stasis, the main alternative explanations for
potential change in the partisan South – the two main causal engines for change
if it came – were not just generally recognized but increasingly invoked in the
relevant professional literature as the post-war era opened, though one did
receive disproportionate attention. Earlier, from the days of Bryan until the
post-war years, agrarian radicalism had been the change engine of choice, and
the Democratic party had seemed the natural theatre for its resulting conflicts.
By the beginning of the post-war period, however, agrarian radicalism was
being displaced by two alternative sources of putative Southern political change.
One of these was legal desegregation and a resulting politics of racial
identification. In this line of argument, it was racial segregation that underpinned
partisan solidarity. Civil rights and a changing politics of race were the great
potential threat to segregation. Partisan solidarity would almost surely crumble
if racial concerns came to the fore. The alternative source of putative partisan
change was economic development and a resulting politics of economic interest.
In this line of argument, it was underdevelopment, as additionally embodied in
a stifling agricultural economy, that underpinned partisan solidarity. Economic
growth and an accompanying politics of class were the great potential
opposition to underdevelopment. Partisan solidarity would almost surely
crumble in their wake. Note that both alternatives, by this time, appeared as
likely to achieve their influence through the Republican as through the
Democratic party.
The foundation stone for both arguments in the immediate post-war years
was, of course, V. O. Key Jr, Southern Politics in State and Nation.
5
It was Key
who most effectively impelled the first great explanation for the distinctive
character of Southern politics – and thus the first great social change that might
ultimately crack it. The central aphorism summarizing his view was destined
to be repeated, often verbatim, in many subsequent studies of Southern politics:
‘Whatever phase of the southern political process one seeks to understand,
sooner or later the trail of inquiry leads to the Negro.’
6
In other words, repressed
4Elections to the United States House of Representatives, 1898–1992, ICPSR/6311 (Ann Arbor:
University of Michigan, Center for Political Studies, 1993); Guide to US Elections (Washington,
DC: Congressional Quarterly Inc., 1975); Congressional Quarterly Weekly Report (Washington,
DC: Congressional Quarterly Inc., 12 November 1994; 6 November 1996; and 7 November 1998).
5V.O. Key Jr, Southern Politics in State and Nation (New York: Alfred A. Knopf, 1949).
6Key, Southern Politics,p.5.
604 SHAFER AND JOHNSTON
racial divisions were the great Southern divide, and whites in the blackest parts
of the South were the critical repressers.
Yet Key himself advanced a second major line of argument about buttresses
for the existing character of Southern politics, and hence about the social change
that might cause them to crumble. Indeed, Key insisted that two great crises, not
one, had shaped and then extended Southern political solidarity: ‘In the second
great crisis whose influence persists – the Populist revolt – political cleavages
often fell along the same lines as in the dispute leading to the War’.
7
Here,
repressed class divisions, that is, divisions in simple wealth and well-being, were
the critical element, though they were sufficiently parallel to repressed racial
divisions as to be practically indistinguishable from them.
In Key’s time, any further effort to compare and contrast these particular
explanations (of the absence of political change) would have been largely beside
the point. A successful one-party resolution was indisputably in place, one that
had obviously triumphed in the face of both latent challenges. Key’s critical
further task was thus to identify the means – the active agents and their practical
strategies – whereby that resolution was maintained:
The almost overwhelming temptation, especially in areas with many Negroes, is
to take advantage of the short-run opportunity to maintain the status quo by using,
or tolerating the use of, the race issue to blot up the discontents of the lesser whites
… with a high degree of regularity those of the top economic groups – particularly
the new industralists – are to be found in communion with the strident advocates
of white supremacy. In the political chaos and demoralization that ensue, alert men
with a sharp eye for immediate advantage take and count their gains …
8
Careful consideration of these same arguments in the light of data unavailable
to Key will largely testify to the accuracy of his arguments. Disciples, however,
were much more attracted by his first line of argument than by his second, as
an engine for the partisan change which would be increasingly apparent to all.
Accordingly, they were to find legal desegregation rather than economic
development, and hence race rather than income, to be the main means for
creating a new resolution, for a new Southern politics:
The newest southern politics, as we have seen, involves three paths to victory. After
the 1994 southern House elections, a much smaller group of white Democrats
survived on the basis of biracial coalitions, an enhanced number of black Democrats
established safe seats based on black majorities, and – for the first time since
Reconstruction – white Republicans controlled a majority of the southern
delegation. The rule of white majorities, the universal road to victory of the white
Republicans, thus reappeared as the central tendency of the newest southern
politics.
9
7Key, Southern Politics,p.7.
8Key, Southern Politics, pp. 662–3.
9Earl Black, ‘The Newest Southern Politics’, Journal of Politics,60(1998), 519–612, p. 607.
The Transformation of Southern Politics, Revisited 605
SOME POLITICAL COMPONENTS OF A REPUBLICAN VOTE
Had he known where he was in history, V. O. Key might have been doubly happy
with Southern Politics in State and Nation.Aspolitical science, the book was
an immediate and then a lasting success. Yet what Key was also doing, by
writing it in the late 1940s, was capturing a portentous historical moment.
Looking backward, he was providing a richly textured description, built around
the essential contours of a partisan world that, if you fuzzed the detail just a bit,
might not have seemed all that different three or even four generations
previously. Looking forward, however, these were the contours that would shift
within one further generation, and shatter within another.
Available indicators of partisan change are numerous. But for an examination
focused through the House of Representatives, nothing can be more central than
the simple progressive capture of House seats by candidates of a Southern
Republican party (see Figure 1a). This aggregate record suggests four general
sub-periods. The record begins with an old world of Southern Democratic
solidarity, born in the nineteenth century, still in place in the 1940s, and
continuing through the 1950s. There is then a sharp, focused breakthrough in
the early 1960s.
10
This is followed by a long, slow, but apparently relentless
assertion of the Republican alternative through the rest of the 1960s, through
the 1970s, and into the 1980s.
11
And there is that final surge to majority status
in the 1990s, leaving a broad array of potential partisan futures in its aftermath.
Fortunately, these sub-periods also suggest that aggregating individual-level
data by decades, a necessity for some subsequent parts of this analysis, will do
no great damage to the historical record, especially if these decades are cut to
begin with the first year when census numbers can actually affect congressional
districts, that is, with 1952, 1962, 1972, 1982 and 1992. Before attempting to
disentangle the impact of racial desegregation and economic development,
however, the partisan record itself can be made to yield some further insights
on Republican fortunes, and thus some further constraints on all subsequent
explanations.
At a minimum, the aggregate record blends (and hence obscures) several
distinguishable strands of the post-war Republican progression. Moreover, the
simplest of these implicit distinctions is ultimately the most powerful, involving
the difference in Republican prospects among:
(1) Democratic seats, the bulk of Southern House seats across this entire period
and nearly the Southern universe when the analysis begins;
10 Alexander Heard, A Two-Party South? (Chapel Hill: University of North Carolina Press, 1952);
Louis M. Seagull, Southern Republicanism (New York: John Wiley, 1975); Alexander P. Lamis, The
Two-Party South (New York: Oxford University Press, 1984).
11 Paul Allen Beck, ‘Partisan Dealignment in the Postwar South’, American Political Science
Review,71(1977), 477–96; John R. Petrocik, ‘Realignment: New Party Coalitions and the
Nationalization of the South’, Journal of Politics,49(1987), 347–75; Harold W. Stanley, ‘Southern
Partisan Changes: Dealignment, Realignment, or Both?’ Journal of Politics, 50 (1988), 64–88.
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
Republican
120
110
100
90
80
70
60
50
40
30
20
10
0
Democrat
Number of wins
(a) Republican seats in the House of Representatives
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
Democratic incumbent
100
90
80
70
60
50
40
30
20
10
0
Republican incumbent*
Mean Republican vote share (percentage)
(b) Open seats and Republican progress
Open seat
* Number of Republican incumbents before 1952 too small for meaningful calculation
606 SHAFER AND JOHNSTON
Fig. 1.The rise of a Southern Republican party
(2) Open seats, those without an incumbent standing for re-election, which rise
and fall intermittently across this period;
(3) and Republican seats, their own incumbent seats, which become a larger
and larger part of the story as the post-war era ages.
The Transformation of Southern Politics, Revisited 607
When this is done, it is clearly the Southern vote for open seats which jumps
up between 1960 and 1962 and then stays on this plateau for the next thirty years
(Figure 1b). Before this jump, the Southern Republican vote for open seats was
derisory, capable of producing victories only when one of the handful of
Republican incumbents departed Congress before the general election.
Afterward, Republicans were entitled to about 40 per cent of the vote in these
open-seat contests, year after year. The result was nothing less than a one-way
escalator, and it became the intermediate vehicle for the Republican progression
in House seats during the thirty years after their breakthrough in the early 1960s.
Where once the Republican party had been entitled to Southern seats only in
long-established bastions, now it was entitled to a substantial minority of seats
whenever they came open.
This thirty-year progression was then capped by a second break and, this time,
that break involved the Republican vote in the face of Democratic incumbents
as well. After 1990, through at least the elections of 1992, 1994 and 1996,
Republican prospects increased sharply, both for picking up open seats and for
challenging and defeating incumbent Democratic congressmen. In 1992, in a
development not widely noted at the time, Republicans actually began to contest
these incumbent seats in a much-improved fashion. In 1994, they improved their
performance with open seats yet again, in a fashion that would prove lasting,
at least as this is written. The combination of those two effects, overlaid on the
long underlying dynamic which had led to this world, was sufficient to produce
a Republican congressional majority in the South.
Yet this final partisan development had the potential to be self-limiting. For
if there were to be more Republican seats and if they were to become safer, then
the Republican vote against Democratic incumbents could potentially fall, as
the Democratic party was increasingly forced back into its own, most secure
base. This seems worth mentioning because it is, at a minimum,
what appeared to happen in 1998, perhaps ending this second concentrated
increment to Republican House seats and perhaps even launching some new
sub-period.
ECONOMIC DEVELOPMENT AND A CHANGING POLITICS OF ECONOMIC
INTEREST
None of that, however, addresses the main cause or causes of the underlying
change, and thus the main explanations for a fresh Republican potential. For this,
it makes sense to turn back initially to the two leading candidates for that status,
both widely touted well before partisan change actually arrived, namely legal
desegregation and economic development. Yet because the Southern electorate
was so overwhelmingly white when the post-war period opened–aregion that
was more than 20 per cent black had a black electorate in the 1950s, after it had
begun to grow, of about 4 per cent – it makes some sense to begin with the white
608 SHAFER AND JOHNSTON
South, and with its divisions, first incipient and then very real, on lines of
economic interest.
The post-war boom for the entire United States was to be historic, with a new
economy, a new occupational structure, and new levels of personal affluence.
Nevertheless, the South managed to narrow the gap with the rest of the nation
on all such measures at the same time.
12
The very basis of its economy was to
change profoundly, from agriculture through manufacturing towards services.
Along the way, a modern occupational distribution inevitably emerged.
13
As a
result, there was a profound change in economic well-being individually, though
this aggregate rise was hardly the whole story: a rising tide did not lift all boats,
and that fact would be central to the post-war partisan evolution of the South.
The main individual-level data for an investigation of the political impact of
economic development, the American National Election Studies (ANES),
present some problems for addressing this effect. Shifting occupational
categories, coupled with shifting means for their assessment, along with
different ways of treating spouses and/or bread-winners, make direct analyses
in terms of social class highly problematic. However, the ANES does categorize
by family income across the whole period from 1952 to 1998. Admittedly, these
are still income bands rather than precise figures. Yet they do allow an easy
division into terciles for the nation as a whole – bottom, middle and top thirds
by income. Moreover, such a division has a special advantage for an
examination of the South: if income terciles are determined for the nation as a
whole, then the South can begin poor (with a disproportionate bottom third) and
grow richer (with a top third growing disproportionately). As in fact it did.
So, it is possible to look at white Southerners for all the post-war decades,
subdivided into income terciles. At the beginning of this period, Southern
Republican congressmen were confined to a handful of Appalachian districts,
which were among the poorest in the entire nation. Their confinement meant that
any class relationship was automatically attenuated. Nevertheless, in the old
world of Southern politics, the relationship which did appear was inverse: that
is, the poor were modestly more likely to vote Republican (Table 1A). That
situation changed, and the engine for a shift toward Southern Republicanism
stood abruptly revealed, in the 1960s. Indeed, from one decade to the next, the
overall relationship actually reversed. A politics of economic interest had
evidently arrived.
12 Jack Bass and Walter DeVries, The Transformation of Southern Politics: Social Change and
Political Consequences Since 1945 (New York: Basic Books, 1976); Black and Black, Politics and
Society in the South; Richard Nadeau and Harold W. Stanley, ‘Class Polarization in Partisanship
among Native Southern Whites, 1952–1990’, American Journal of Political Science,37(1993),
900–19; Bartley, The New South.
13 John C. McKinney and Linda Brookover Bourque, ‘The Changing South: National
Incorporation of a Region’, American Sociological Review,36(1971), 399–412; James C. Cobb,
Industrialization and Southern Society, 1877–1984 (Lexington: University of Kentucky Press, 1984);
James C. Cobb, The Selling of the South: The Southern Crusade for Industrial Development,
1936–1990 (Champaign-Urbana: University of Illinois Press, 1993).
The Transformation of Southern Politics, Revisited 609
TABLE
1Social Class and Southern Republicanism
Income terciles Range
(high- Decadal
Decade Low Middle High Total low) change
A. Republican percentage among all whites by income level
1950s % Republican 18% 15% 15% 16% ⫺3% –
Total Votes (N) (199) (139) (222) (560) (560)
1960s % Republican 25% 28% 35% 29% ⫹10% ⫹13%
Total Votes (N) (223) 208) (224) (655) (655)
1970s % Republican 24% 34% 38% 32% ⫹14% ⫹4%
Total Votes(N) (270) (312) (287) (869) (869)
1980s % Republican 27% 37% 49% 39% ⫹22% ⫹8%
Total Votes (N) (226) (300) (304) (830) (830)
1990s % Republican 40% 61% 70% 59% ⫹30% ⫹8%
Total Votes (N) (216) (247) (324) (787) (787)
B. Republican percentage among all whites by income level in districts with both
Democratic and Republican candidates
1950s % Republican 40% 32% 25% 31% ⫺15% –
Total Votes (N) (81) (60) (102) (243) (243)
1960s % Republican 31% 37% 44% 37% ⫹13% ⫹28%
Total Votes (N) (137) (113) (122) (372) (372)
1970s % Republican 33% 44% 50% 43% ⫹17% ⫹4%
Total Votes (N) (183) (215) (195) (593) (593)
1980s % Republican 35% 51% 54% 48% ⫹19% ⫹2%
Total Votes (N) (125) (196) (219) (540) (540)
1990s % Republican 38% 57% 63% 54% ⫹25% ⫹6%
Total Votes (N) (185) (202) (236) (623) (623)
Or at least, in the 1960s, the wealthiest tercile was most likely to vote
Republican, the poorest tercile least likely. The basic direction of the
relationship had changed, in a turn which was never in any way threatened
thereafter. And while the move from the 1950s to the 1960s was also the
strongest overall shift, all subsequent decadal changes were to move further in
the same direction. The 1950s, as captured by these individual-level data in
tabular form, is the same world that Key had captured with ecological data
graphed onto maps. The 1960s is not. On the one hand, this will prove to be the
overall story of partisan change among whites in its relationship to personal
income during all the post-war years in the American South. On the other hand,
its aggregation still masks the extent of the change. For it still includes many
who could not vote Republican for Congress, because they had no Republican
candidate.
Accordingly, the analysis should really be restricted to white Southerners who
at least possessed both a Democratic and a Republican congressional candidate.
610 SHAFER AND JOHNSTON
When it is, previous patterns return, writ larger still (Table 1B). Now, for the
1950s, in those districts where there actually was a Republican candidate, the
economic relationship to Republican voting stands sharply inverted. Forty per
cent of the poorest tercile, a third of the middle tercile but just a quarter of the
top tercile votes Republican. The 1960s again reverse this relationship, more
dramatically this time. The share of Republican voters in the top third of incomes
nearly doubles; the share of Republican voters in the bottom third simul-
taneously falls. Moreover, the reversal between this relationship and its
counterpart in the 1950s remains by far the largest such shift anywhere in these
data. That relationship is then extended in the 1970s. Republican support is
ratcheted up further in the 1980s. And the second largest jump, the main
secondary surge, occurs in the 1990s.
Already in the 1970s, however, a majority of the top income tercile was voting
Republican for Congress when these individuals had the chance to do so. In the
1980s, that majority advanced and a majority of the middle tercile began to vote
Republican. In the 1990s, as this is written, nearly two-thirds of the top tercile
cast a Republican vote. This is probably not the sharpest economic cleavage in
American political history.
14
But it is an impressively sharp cleavage for a
politics usually argued to be shifting towards other bases elsewhere in the United
States during this same period.
15
Several other aspects of this story of the coming of income divisions to the
American South – and perhaps of that underlying class politics already typical
of the American North – seem worth noting (Table 1B). From one side, the top
income tercile moved regularly and relentlessly upward, from an old base of 25
per cent Republican in the 1950s to a current peak of 63 per cent Republican
in the 1990s. From the other side, the bottom tercile did not move up at all, and
this fact probably deserves emphasis: in a period in which the partisan loyalties
of the entire region were being overturned – the greatest pro-Republican shift
in the South since Reconstruction had imposed one – the bottom third of white
society did not join this movement, and in fact drifted slightly more Democratic.
LEGAL DESEGREGATION AND A CHANGING POLITICS OF RACIAL
IDENTIFICATION
One reason to begin an analysis of the composition of post-war partisan change
with white voters, subdivided by income, is that at the beginning of this period
14 David Montgomery, The Fall of the House of Labor: The Work Place, the State, and American
Labor Activism,1865–1925 (New York: Cambridge University Press, 1987); Stuart M. Blumin, The
Emergence of the Middle Class: Social Experience in the American City, 1760–1900 (New York:
Cambridge University Press, 1989).
15 Everett Carll Ladd with Charles D. Hadley, Transformations of the American Party System
(New York: W.W. Norton, 1975); Byron E. Shafer, Quiet Revolution: The Struggle for the
Democratic Party and the Shaping of Post-Reform Politics (New York: Russell Sage Foundation,
1983); Allen J. Matusow, The Unraveling of America: A History of Liberalism in the 1960s (New
York: Harper & Row, 1984).
The Transformation of Southern Politics, Revisited 611
TABLE
2Racial Background and Southern Republicanism: Republican
Percentage among Black Southerners in Districts with Both
Democratic and Republican Candidates
Income terciles
Decade Low Middle/High Total
1950s % Republican 50% 0% 40%
Total Votes (N) (4) (1) (5)
1960s % Republican 1% 1% 1%
Total Votes (N) (37) (22) (59)
1970s % Republican 1% 12% 4%
Total Votes (N) (64) (34) (98)
1980s % Republican 1% 1% 1%
Total Votes (N) (86) (64) (150)
1990s % Republican 1% 1% 1%
Total Votes (N) (85) (61) (146)
there was little point even inquiring into the voting behaviour of Southern
blacks. Diagnostically, the ANES surveys for the 1950s, all together, offer a
pooled total of exactly two votes for Republican congressional candidates by
black Southerners, so that the survey base for any further analysis is exiguous.
Yet change in the racial organization of Southern society, too, was about to
occur.
16
Registration by Southern blacks accelerated remarkably: 5 per cent in
1940, 20 per cent in 1952, 29 per cent in 1960, but 65 per cent by 1969.
17
Voting
by Southern blacks followed inexorably: 4.1 per cent of the Southern electorate
in the 1950s, 12.3 per cent in the 1960s, 16.0 per cent in the 1970s, 21.8 per cent
– their proportionate share – in the 1980s.
18
As it developed, however, there was still no point in inquiring into the
economic politics of this newly enfranchised black electorate, albeit for a
different reason. If it was their own regional Democrats who had most strongly
resisted bringing black Southerners into the formal electorate, it was national
16 Numan V. Bartley and Hugh Davis Graham, Southern Politics and the Second Reconstruction
(Baltimore, Md.: Johns Hopkins University Press, 1975); Bass and DeVries, The Transformation of
Southern Politics; Harvard Sitkoff, The Struggle for Black Equality, 1954–1980 (New York: Hill
& Wang, 1981); Hugh Davis Graham, The Civil Rights Era: Origins and Development of National
Policy, 1960–1972 (New York: Oxford University Press, 1990).
17 J. Morgan Kousser, The Shaping of Southern Politics: Suffrage Restriction and the
Establishment of the One-Party South, 1880–1910 (New Haven, Conn.: Yale University Press,
1974); Bartley and Graham, Southern Politics and the Second Reconstruction; Steven F. Lawson,
Black Ballots: Voting Rights in the South, 1944–1969 (New York: Columbia University Press, 1976).
18 Derived from Tables 1 and 2. See also Donald R. Matthews and James W. Prothro, Negroes
and the New Southern Politics (New York: Harcourt, Brace & World, 1966); William R. Keech, The
Impact of Negro Voting: The Role of the Vote in the Quest for Equality (Chicago: Rand McNally,
1968); Harold W. Stanley, Voter Mobilization and the Politics of Race: The South and Universal
Suffrage, 1952–1984 (New York: Praeger, 1987).
612 SHAFER AND JOHNSTON
Democrats who had most strongly supported their arrival, and black Southerners
did not fail to notice: they were to be an overwhelmingly Democratic bloc from
the moment of their arrival, and thus a hugely cohesive bloc by definition. As
a result, there would be little room for further internal analysis, on economic
or, evidently, any other grounds (Table 2). Beginning as a desperately poor
social group, black Southerners must, for analytic purposes, be divided merely
into a bottom economic third versus the rest. Yet even in this simple division,
the resulting partisan situation is abundantly clear.
When black Southerners are considered (like their white counterparts) in only
those districts which offered the opportunity to vote either Democratic or
Republican, there is simply no income relationship at all (Table 2). Or at least,
from the point at which the Voting Rights Act kicks in, during the 1960s, black
voters are overwhelmingly likely to vote Democratic for Congress, decade in
and decade out, bottom tercile or not. Numbers like that made internal
distinctions unimportant, and Republican votes derisory. The Southern black
vote, brought into the electorate by indigenous racial agitation and national
racial legislation, was a vote based in racial identity, not economic interest.
However, that was not quite the end of the story of an evolving racial politics
for the post-war South. Economic development did release the genie of
economic interest into active Southern politics, and income became hugely
influential among whites, providing a rationale on its own for a new Southern
white Republicanism. Desegregation then kept racial identity at the centre of this
same politics, albeit in a different way: race blanketed income among black
Southerners, effectively trivializing it. Yet this did not in principle prevent a
class-based white Southern politics from being additionally influenced by
considerations of race, at least in one simple regard. Indeed, many students of
Southern politics long felt compelled – felt that the anecdotal record obviously
and inescapably justified – attending to matters of ‘racial threat’.
19
As indeed
it does.
Because a sizeable black minority, newly added to partisan politics, would
automatically require sizeable adjustments within the dominant Democratic
party, it might constitute a kind of second-order opportunity for aspiring
Republicans as well. If new black voters were overwhelmingly Democratic, old
white Democrats might feel themselves displaced. Below a certain level, both
the theoretical argument and the practical opportunity remained trivial. Yet the
South as a whole was almost exactly 20 per cent black in the 1970 census, the
one after the voter registration gains of the 1960s. And a bloc of more than 20
per cent of the potential partisans in a district would effectively demand some
consideration – some conciliation or some confrontation – even if that district
remained overwhelmingly Democratic. Obviously, a minority bloc of that size
would loom only larger if a Republican party became a serious alternative,
especially since this particular bloc, unlike many others, was not divided by
19 Michael W. Giles, ‘Percent Black and Racial Hostility: An Old Assumption Revisited’, Social
Science Quarterly,70(1977), 820–35; Black, ‘The Newest Southern Politics’.
The Transformation of Southern Politics, Revisited 613
cleavages (like income) that would otherwise give it its own, internal, partisan
divisions.
It is not difficult to isolate white voters once again, this time according to
whether they lived in congressional districts that were more or less black than
the Southern average. When this is done, a clear-cut story does emerge (Table
3). If one looks first at the white South by racial composition only, ignoring for
the moment those strong class divisions among white Southerners – that is,
looking only at the right-hand columns in Table3–the pattern is clear: whites
in districts with a larger black population were consistently less likely to vote
Republican than whites in districts with a less consequential black minority. It
is not clear how much attention should be paid to this relationship in the 1950s,
when disfranchised blacks could have constituted only a very different kind of
‘threat’ to white voters than they would become when they could actually
register and vote. But for the 1960s, the 1970s and the 1980s, the aggregate
relationship remains clear and consistent: the blacker the district, the less likely
whites were to vote Republican. Only in the 1990s did this relationship finally
fade, to effective racial neutrality.
Yet the very strength of income-based considerations within this white
electorate, at least from the 1960s onward, does raise the possibility that this
race-based relationship is itself an artefact. Outside Appalachia, districts with
more blacks were also poorer districts, and districts with fewer blacks were
wealthier, so that this first cut at the impact of racial context might always reflect
just economic development and income politics. In actuality, however,
apparently not (Table 3). For the 1960s and 1970s and 1980s, in every
comparison – low-income whites in more and less black districts, middle-in-
come whites in more and less black districts, or upper-income whites in more
and less black districts – every category of white voter was less likely to vote
Republican in the side of this comparison that included more black voters. This
was hardly evidence of more (or less) racial tolerance: these data are agnostic
on that. But it was an evidently robust pattern to white voting by racial context.
20
LEGAL DESEGREGATION AND THE POLITICS OF ECONOMIC INTEREST
Those are two main and direct contributions to Republican prospects from
structural change in the post-war South. Both are large, and if the coming of a
refreshed economic politics among white Southerners was ultimately larger,
slowly but inexorably creating a Republican House delegation in the Old
Confederacy, the coming of a refreshed racial politics among black Southerners
20 The place of racial threat in that account has now come under increasing attack, even on its
own terms. See Alan I. Abramowitz, ‘Issue Evolution Reconsidered: Racial Attitudes and
Partisanship in the US Electorate’, American Journal of Political Science,38(1994), 1–24; and
especially Stephen D. Voss, ‘Beyond Racial Threat: Failure of an Old Hypothesis in the New South’,
Journal of Politics,58(1996), 1156–70.
614 SHAFER AND JOHNSTON
TABLE
3Racial Contexts and White Voters: Republican Percentage among All Whites by Racial Composition of
Districts with Both Democratic and Republican Candidates
Income terciles
Low Middle High Total
District District District District District District District District
⬍20% ⬎20% ⬍20% ⬎20% ⬍20% ⬎20% ⬍20% ⬎20%
Decade Black Black Black Black Black Black Black Black
1950s % Republican 43% 23% 46% 12% 19% 32% 35% 24%
Total Votes (N) (68) (13) (35) (25) (58) (44) (161) (82)
1960s % Republican 37% 20% 54% 22% 47% 38% 45% 27%
Total Votes (N) (73) (60) (48) (63) (57) (61) (178) (184)
1970s % Republican 39% 25% 51% 34% 53% 46% 48% 35%
Total Votes (N) (114) (69) (123) (92) (123) (72) (360) (233)
1980s % Republican 44% 20% 62% 31% 68% 35% 60% 31%
Total Votes (N) (80) (45) (122) (74) (125) (94) (327) (213)
1990s % Republican 37% 40% 57% 60% 61% 69% 53% 56%
Total Votes (N) (145) (40) (160) (42) (194) (42) (499) (124)
The Transformation of Southern Politics, Revisited 615
was more encompassing for those whom it affected. Yet these two great
structural changes also interacted, in three further ways.
The first may be self-evident. If the coming of economic development, and
with it an ‘income escalator’ for Republican prospects, is the main story of a
Republican South; but if the coming of legal desegregation, and with it a huge
influx of new, Southern, black Democrats, is the main secondary story; then
obviously the latter served as a major restraint on the former. Had relationships
between racial context and economic interest among white voters been different
– that is, had greater black percentages coincided with greater white
Republicanism – then this effect would have been reinforcing rather than
countervailing, and the story of net impact would have been quite different as
well. But it evidently was not.
Instead, the scale of the actual combined effect was so substantial as to be
worth some additional comment. Tables 1b and 2 have implicitly summarized
this aspect of the story, and the results in their columns marked ‘Total’ can be
re-presented and used here merely to underline its importance:
1950s 1960s 1970s 1980s 1990s
White Southerners only 31% 37% 43% 48% 54%
Entire Southern electorate 31% 32% 37% 38% 44%
Now, the overall damping effect of legal desegregation on economic
development, as a prod to partisan change, stands out additionally. If Republican
prospects moved relentlessly up the economic escalator after the 1950s, a major
influx of black Democrats also held those prospects back at every point. Even
in recent years, even in the period when Republican voting gains were
increasingly and directly reflected in Republican congressional wins, the civil
rights revolution was still holding back the Republican tide. In the absence of
that revolution (as in the row for ‘White Southerners only’), the votes of white
Southerners might have been sufficient to deliver the dramatic outcome of 1994
in the mid-1980s instead!
That was a straightforward countervailing effect: race against economics. Yet
there was a major reinforcing effect as well; ironically, it too lowered the line
of Republican progress. It should be no surprise that the coming of black
Southerners into the mass electorate after the civil rights legislation of the 1960s
altered the economic composition of the total electorate. This legislation was
focused, after all, on a desperately poor segment of American society. But in
fact, the elimination of barriers to registration and voting swelled the white
electorate, by itself, and this too altered the economic composition of the
political South in an anti-Republican way.Ifweagain go immediately to whites
with a full partisan choice, those possessing both Democratic and Republican
congressional candidates for whom potentially to vote, the story surfaces starkly
(Table 4).
Seen this way, the edge for the top income tercile of white Southerners in the
electorate of the 1950s had been truly disproportionate, given how underdevel-
oped the South was overall. The civil rights revolution then sharply decreased
616 SHAFER AND JOHNSTON
TABLE
4Economic Development, Legal Desegregation,
and the Republican Vote: Composition of the
White Southern Electorate by Social Class for
Districts with Democratic and Republican
Candidates
Income terciles
Decade Low Middle High N
1950s 0.33 0.25 0.42 243
1960s 0.37 0.30 0.33 372
1970s 0.31 0.36 0.33 593
1980s 0.23 0.36 0.41 540
1990s 0.30 0.32 0.38 623
the electoral share of this wealthiest Southern third in the 1960s, while the
middle tercile rose and the poorest tercile actually became the modal category
of southern white voters. In the 1970s, the middle tercile seized this modal role
instead. And by the 1980s, the top tercile had reclaimed its predominance –
apparently reasserting through economic growth and voter turnout, three
decades later, what it had once possessed through advantageous rules. If the
situation appears less extreme in the 1990s, this decline is largely an artefact of
the increase in the share of uncontested Republican seats, where top-tercile
whites are especially concentrated.
Those are two quite different effects of the interaction of race and class. What
unites them, nevertheless, is the fact that each contributed a substantial restraint
on the line of Republican progress. The last such interaction, however, one
where racial politics finally trumps economic politics, actually constituted an
important increment to the Republican successes of the 1990s, the successes
which were crucial in finally bringing the Republican party to majority status
in the South. In the process, a race effect finally augmented the underlying
economic dynamic.
Yet the 1990s were also the decade of the ‘minority majority’ district, of
congressional districting aimed at increasing the presence of black congressmen
and women from the South.
21
In the 1960s, before the civil rights revolution
made the numbers relevant, there were two congressional districts with a black
majority in the entire South; by the 1980s, there were still only three; in the
1990s, by design, there were seventeen. It has become commonplace to note that
the process of creating these districts is also a process of advantaging Republican
21 Kevin A. Hill, ‘Does the Creation of Majority Black Districts Aid Republicans? An Analysis
of the 1992 Congressional Elections in Eight Southern States’, Journal of Politics,57(1995),
384–401; David Lublin, The Paradox of Representation: Racial Gerrymandering and Minority
Interests (Princeton, NJ: Princeton University Press, 1997).
The Transformation of Southern Politics, Revisited 617
candidates overall. As reliable Democrats (which black voters certainly are)
become concentrated in a smaller number of districts, the potential Republican
share of the remainder must rise.
Table 3, however, suggests an additional effect. Or at least, if white residents
are more likely to vote Democratic in districts with a higher share of black
voters, then the process of moving white voters out of those districts – which
is also what minority-majority districts do – should not just shuffle white
Democrats from one district to another. It ought additionally, or so these data
suggest, to make them more Republican in the process. The specifics of this
effect are hard to isolate with these data. Before 1990, there were only three
districts with black majorities; after 1990, there were few whites left in districts
with disproportionate black populations. But the overall direction seems
inescapable, and the jump from three to seventeen districts would make it
additionally consequential.
22
THE APPEARANCE OF CHALLENGERS AND THE RESISTANCE OF
INCUMBENTS
So far, this has been an analysis of structural shifts and partisan change, where
major elements of social structure drive partisan outcomes in electoral politics.
Yet there were crucial intermediaries to this translation, in the form of political
parties. Their leading figures may not have pictured themselves as critical
buffers between economic development and legal desegregation. But they did
see changes on the ground and they did respond aggressively, either to cope with
or to capitalize upon those shifts. In the process, they contributed – they became
– major explanations for both the pace and the specifics by which structural
shifts elicited partisan change.
The presence of a sitting incumbent remained a powerful brake on Republican
prospects across almost this entire period, even when it failed to deter a
Republican candidacy in opposition.
23
Indeed, even with seats that offered no
Republican challengers excluded from the analysis, sitting Democratic
incumbents remained remarkably successful at repressing the Republican vote.
(Figure 2a) There was a modest uptick to this Republican tally in the face of
incumbent Democrats after 1960, but it was truly modest. More striking was the
22 A very promising attack on the problem, with parallel implications, is David Lublin and
Stephen D. Voss, ‘Racial Redistricting and Realignment in Southern State Legislatures’ (paper
presented at the Annual Meeting of the American Political Science Association, Atlanta, Georgia,
1999).
23 David W. Rohde, ‘ “Something’s Happening Here; What It Is Ain’t Exactly Clear”: Southern
Democrats in the House of Representatives’, in Morris P. Fiorina and Rohde, eds., Home Style and
Washington Work: Studies of Congressional Politics (Ann Arbor: University of Michigan Press,
1991); Nicol C. Rae, Southern Democrats (New York: Oxford University Press, 1994); M. V. Hood
III, Quentin Kidd and Irwin L. Morris, ‘Of Byrd [s] and Bumpers: Using Democratic Senators to
Analyze Political Change in the South, 1960–1995’, American Journal of Political Science,43
(1999), 465–87.
618 SHAFER AND JOHNSTON
stability of the Republican vote, at a low level, whenever a Democratic
incumbent stood for re-election. This line moved hardly at all across all the
post-war years. Which is to say: the great intermediary reason that the line of
Republican progress from 1962 through 1992 was a modest one, albeit
relentlessly upward, was the broad and general presence of Democratic
incumbents – and their powerful, individual and collective, damping effect.
The situation was strikingly different in the case where the ‘Grand Old Party’
(the Republicans) produced a challenger and there was no incumbent Democrat
(the other line in Figure 2a). The average Republican vote in the absence of a
sitting Democrat was always considerably higher than the average vote for such
a Republican challenger in the face of an incumbent. Yet this gap actually
increased during the long upward march from 1962 through 1990. In this regard,
it is worth noting additionally that the vote for Republican challengers to
Democratic incumbents was of some consequence earlier than the period of
substantial Republican victories – that is, in the years before 1962 – suggesting
that an incipient potential may have been present even earlier than the won/lost
tally would indicate.
Incumbency, of course, had a second, indirect effect on the appearance of a
Republican alternative. In this feed-back effect, not only were Republican
challengers likely to do much less well against Democratic incumbents than in
open seats. They were also less likely to appear – there were less likely to be
Republican challengers – when there was a sitting Democrat, bringing the two
effects back together (Figure 2b). In the years before 1962, Republican
challengers were a rarity when there was a Democratic incumbent, and relatively
rare even when there was not. The propensity/ability to provide challengers was
already rising sharply in open seats by the mid-1950s – there were fewer and
fewer unchallenged opportunities of this sort – again suggesting that an apparent
potential was taking shape below the surface of politics.
In any case, by 1962 and then ever after, whenever there was no sitting
Democrat, there was almost sure to be a Republican challenger for a Southern
seat in the US House: the share of unchallenged open seats fell under 10 per cent
for 1962 and did not rise again to that level. However, the line for Republican
challenges to Democratic incumbents was even more striking (Figure 2b). In the
1950s, it was possible to provide such challenges in about a quarter of all
relevant districts. Between 1962 and 1990, it was possible to provide them in
roughly half the total opportunities, with a modest improvement as the years
passed. Then in 1992 and for the next two elections, the share of unchallenged
Democratic incumbents fell effectively to zero as the Southern Republican
majority arrived. Finally, in 1998, this share bounced back sharply, as both
Southern Republicans and Southern Democrats turned to concentrating their
resources on a potential new pattern of party competition.
In a long-time one-party region, where Democratic incumbents were
omnipresent and where sporadic Republican candidacies had been effectively
hopeless outside of party strongholds left over from the days of the Civil War,
the provision of fresh Republican candidates was obviously essential to
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
Districts with a Democratic incumbent
90
80
70
60
50
40
30
20
10
0
Districts without a Democratic incumbent
Mean Republican vote share (percentage)
(a) Fate of Republicans challengers by Democratic incumbency
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
Districts with a Democratic incumbent
100
90
80
70
60
50
40
30
20
10
0
Percentage of races with a Republican candidate
(b) Appearance of Republican challengers by Democratic incumbency
Districts without a
Democratic incumbent
The Transformation of Southern Politics, Revisited 619
Fig. 2.Incumbent status and Republican growth
capitalizing on any incipient voting potential, any putative partisan demand
arising from underlying structural change. Not surprisingly, then, the record of
candidate provision also tracks the overall record of Republican successes
(Figure 3a). There is the same sharp uptick in the addition of Republican
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
90
80
70
60
50
40
30
20
10
0
Democrats
Percentage of districts contested
(a) Growth in Republican contestation of seats
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
All districts
70
60
50
40
30
20
10
0
Mean Republican vote share (percentage)
(b) Republican candidacies and moblilization of the vote
Districts with a
Republican candidate
100
Republicans
620 SHAFER AND JOHNSTON
Fig. 3.Candidate provision and Republican prospects
candidates after 1960, to which the presidential candidacy of Barry Goldwater
added impetus.
24
This is followed, once again, with a long, slow, but ineluctably
upward crawl of Republican candidacies from 1966 to 1990. And then comes
24 Theodore H. White, The Making of the President, 1964 (New York: Atheneum Publishers,
1965), chaps. 3–5; Nicol C. Rae, The Decline and Fall of the Liberal Republicans: From 1952 to
the Present (New York: Oxford University Press, 1989), chap. 2.
The Transformation of Southern Politics, Revisited 621
that other substantial break, in 1992 and after, when the conscious efforts of the
Republican National Committee – its belief that remaining votes were there to
be harvested – again added impetus to the background trends.
25
By the time this final push was under way, the Republican party was fully
competitive across the South in the most fundamental way possible: it had
Republican congressional candidates everywhere. By 1994, in fact, it had more
such candidates than the Democrats. The appearance of challengers is thus a
further and important part of the dependent variable, of the Republican vote and
Republican successes.
26
The recruitment of these challengers allowed the party
to harvest an incipient base of support.
27
Nevertheless, they hardly constituted
that base: in the absence of some other change creating a ‘pent-up’ Republican
potential, a greater number of challengers would only have lowered the average
vote for Republican candidates. But in fact, across all the years after 1960, they
did not (Figure 3b).
In fact, the social changes contributing this Republican potential affected
even the appearance of Republican challengers, and in major ways. Moreover,
this impact provides additional confirmation that a focus on districts with both
Democratic and Republican candidates is not obscuring some other, major part
of the story. The Bureau of the Census did not break down occupational statistics
by congressional district for the 1950s. But by 1962, when the Republicans
began their inexorable rise, the party was effectively able to provide candidates
in almost all districts to which the Census assigned a white-collar majority
(Figure 4a). There were still very few such districts in the 1960s. Yet the
appearance of Republican candidates across the post-war years was to ride up,
essentially, on the gradual growth of the white-collar South. Districts with
blue-collar majorities drew candidates in a considerably lower proportion of
cases, a proportion which increased only very gradually – until the 1990s, when
these became the districts dragged up by conscious effort.
The racial composition of districts was far less important here, though it too
ran in the expected direction (Figure 4b). On the one hand, provision of
candidates in more-black districts tracked provision of candidates in less-black
counterparts from the 1950s until the 1950s. On the other hand, less-black
districts always got somewhat more attention until that time. Regardless, for
districts categorized either way, by income or by race, this pattern of candidate
appearance also confirms that a focus on districts which did possess both a
Democratic and a Republican candidate is not distorting the analysis. From the
25 William F. Connelly and John J. Pitney, Congress’ Permanent Minority? Republicans in the
U.S. House (Lanham: Rowman & Littlefield, 1994); Walter Oleszek, ed., The 104th Congress: A
Congressional Quarterly Reader (Washington, DC: CQ Press, 1995).
26 David R. Mayhew, Congress: The Electoral Connection (New Haven, Conn.: Yale University
Press, 1974); Richard F. Fenno Jr, Home Style: House Members in Their Districts (Boston, Mass.:
Little Brown, 1978).
27 Gary C. Jacobson and Samuel Kernell, Strategy and Choice in Congressional Elections (New
Haven, Conn.: Yale University Press, 1981); Linda L. Fowler, Candidates, Congress, and the
American Democracy (Ann Arbor: University of Michigan Press, 1993).
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
90
80
70
60
50
40
30
20
10
0
Districts more
than 50%
white-collar
Percentage of districts with Republican candidate
(a) Republican candidacies by class composition of district
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
Districts more
than 20%
black
70
60
50
40
30
20
10
0
Percentage of districts with Republican candidate
(b) Republican candidacies by racial composition of district
Districts
less than 20%
black
100
Districts
less than 50%
white-collar
80
90
100
622 SHAFER AND JOHNSTON
Fig. 4. Structural change and candidate provision
1960s, as districts are added over time to the total roster of those which do
possess candidates from both parties, the relationship between income or racial
context and Republicanism does not alter.
The Transformation of Southern Politics, Revisited 623
OLD SOUTH
,
NEW SOUTH
,
NO SOUTH?
That is a story of the evolution of post-war politics in the American South. From
at least the 1960s onward, there was an underlying partisan dynamic to Southern
politics, an implicitly Republican dynamic fuelled by economic development,
and it ran in regular and relentless fashion until it had accumulated a majority
of Southern seats in the US House of Representatives. By contrast, the great
barrier to this dynamic was the civil rights revolution, both directly and
indirectly. Moreover, until very late in the day, that dynamic could break
through only when Democratic incumbents removed themselves from the
contest – and an established generation of senior Southern congressmen
managed to live a very long time in such an environment. Yet whenever
economic development met incumbent retirement, in truth as early as the 1950s,
then a Republican replacement was the likely outcome.
At first, economic development and its attendant white-collar majorities were
themselves in short supply. But come they ultimately did, with a vengeance,
while even unchallenged incumbents eventually went. The resulting dynamic
was as simple as it was inescapable. Economic development produced a
changing politics of economic interest; economic politics produced a Republi-
can majority. By contrast, racial solidarity remained as a central aspect of
southern politics. Moreover, it remained a central aspect for the same party, the
Democrats.
Yet the civil rights revolution also came, so that where once racial solidarity
was a distinguishing characteristic of the white South, helping to deny any
foothold for an economic politics, now it was a distinguishing characteristic of
the black South instead. In the end, then, the two Southern parties, at least in
office in the House of Representatives, looked remarkably like a reversal of their
previous selves. Where once there was a minority party based on concentrations
of poor whites plus blacks everywhere, now there was still a minority party
based on concentrations of poor whites plus blacks everywhere. It was just that
once this party had been the Republicans, and now it was the Democrats.
In many ways, there is no natural end to this story. Trend lines which had no
inevitability in the past should be given no inevitability for the future. Moreover,
it is probably worth noting how much of the Southern story still lies outside the
confines of this article. We have not added the attitudinal dimension to the tale
told here; we have not added struggles for other offices, most especially the
presidency; we are not, with the partial exception of the next few paragraphs,
scratching the surface of any South/North comparison. We are at work on all
three, but they must be subjects for subsequent papers. Nevertheless, there is
one major aspect of the story which appears to have reached closure. For what
we have probably been observing in all of the preceding is, at a higher level of
abstraction, a huge instance of the nationalization of partisan politics. Or at least,
a handful of closing North/South comparisons does suggest that such a
nationalization has finally arrived.
The leading elite indicator, the appearance of Republican candidates, can
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
90
80
70
60
50
40
30
20
10
0
Non-South
% of districts with Rep.candidate
(a) Republican congressional candidacies by region
100
South
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
60
50
40
30
20
10
0
Non-South
Mean Rep. vote share (percentage)
(b) Republican congressional votes by region
South
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
80
70
60
50
40
30
20
10
0
Non-South
% of seats won by Rep. candidates
(c) Republican congressional seats by region
South
624 SHAFER AND JOHNSTON
Fig. 5. The nationalization of the party system
The Transformation of Southern Politics, Revisited 625
easily be compared (Figure 5a). The non-Southern picture is effectively a
straight line, with a bit of hesitation in the 1980s but otherwise with nearly
complete and uniform provision of Republican candidates. The Southern picture
is the familiar one of distinct sub-pieces. There is an old world, for the 1950s
and before, of candidate provision in less than half of all House districts. There
is a changed world, of surge in the early 1960s and then further growth through
the 1980s, when about three-quarters of all Southern districts could expect to
have a Republican House candidate. And there is a new world, from the early
1990s, when nearly every district could. As a result, this is also a story of
convergence in candidate provision, South and non-South.
The leading mass indicator, the actual Republican vote by region, does not
just tell another part of the same story. It testifies to the same regional
convergence (Figure 5b). Here, the non-South shows a long period of narrow
oscillation within the Republican vote for the US House, at between 40 and 50
per cent of the total. The South instead shows its same sub-periods: a derisory
vote through the 1950s, a break in the early 1960s with a gradually encouraging
increase through the 1980s, and a jump to majority status in the 1990s. By then,
once more, South and non-South had effectively converged, with Southern
Republicans performing – mirabile dictu –abit better.
Last but not least – last but most consequential, actually – is the indicator in
which elite provision and mass response come together, namely the final record
of wins and losses, of public offices gained, public offices surrendered and
public offices held (Figure 5c). There is little new to say about it. After the very
beginning of the Eisenhower interregnum, non-Southern Republican prospects
for the US House settled into that same band of 40–50 per cent for most of the
post-war years. By contrast, Southern prospects were remarkably derisory into
the 1960s. They rose relentlessly from there through the 1980s. And they jumped
to majority status in the 1990s.
When they did, they had effectively converged on the same line as the
non-South. Whether or not they stay there – there are several ways in which these
lines can go on to diverge again, and the election of 1998 was an especial tease
here – a new institutional era for the parties, most especially in the House of
Representatives, appeared to be at hand. And the South appeared to be its main
story.
625a SHAFER AND JOHNSTON
APPENDIX: A NOTE ON DATA ANALYSIS
Three principal bodies of data were merged to permit this analysis (see the introductory
section), and the resulting merged dataset was sufficiently comprehensive to facilitate it.
Nevertheless, there were three intermittent anomalies remaining within that dataset, each
requiring some further decision-rule:
— The first involved court-ordered, mid-decade reapportionments. These were most
consequential in the 1990s, though we adjusted district demographics to comply with them
in every case. A few of the resulting calculations were not precise to the individual, but none
could have resulted in the misclassification of a district in any of the preceding tables or
figures.
— The second of these decision-rules involved at-large elections to the House, of which there
were a sprinkling before 1970. This resulted when a state failed to reapportion in time for
the first election after a new national census, and was thus forced to conduct that election
at-large for any new district. Because percentages, outcomes and incumbency (rather than
aggregate votes) were central to our tables and figures, these numbers could still be used,
as with any other district.
The only larger instance came in Alabama in 1962, when the state conducted all eight
of its congressional elections at-large. When this did not present a problem for the preceding
tables and figures, as with the number of Republican seats, we included Alabama 1962.
When it did create a problem, as with Republican vote for open seats, we treated it as missing
data. Note that this constituted only eight of 530 Southern congressional districts for the
1960s, even then.
— The third decision-rule was necessitated by the adoption of a unified primary in the state
of Louisiana from 1976 onward. In this arrangement, the top two finishers (regardless of
party) contest the general election, unless one of them acquires an outright majority in the
primary. Where a Democrat ran against a Republican in the general election – the most
common outcome when the election was contested – this was the vote used in the dataset.
Because our focus is on the rise of a Republican party and a Republican vote, however, where
two members of the same party ran against each other at the general election but where there
were both Democratic and Republican candidates in the primary – only a handful of cases
– the aggregated vote for Democrats and Republicans at this earlier stage was used instead.
