Attempts to Crack the Case of the Vanishing Marginals
- Zayne Sember
- Dec 8, 2019
- 11 min read
In his 1974 “The Case of the Vanishing Marginals,” as the Sherlock Holmes-esque title suggests, David Mayhew investigates a phenomenon that he terms the “vanishing marginals” – a precipitous decline in US House elections with close margins. These vanishing marginals are summarized by Figure 1, showing the number of districts with incumbents running with both a narrow (left) and wide (right) definition of a marginal election. For both definitions, a marked decline is visible – tight House races appear to be vanishing. This simple observation sparked a subsequent flurry of analysis of and debate about close House elections by political scientists in the following decades as the causes and implications of the vanishing marginals were investigated.
Unlike Sherlock Holmes, Mayhew does not attempt to crack the case of the vanishing marginals, saying: “Why the decline in incumbent marginality? No clear answer is available.” (Mayhew 2008, 62). His only offers to explain the vanishing marginals are five hypotheses. The first is the line-drawing explanation – the idea that gerrymandering creating safer districts is to blame. However, despite being a frequent explanation offered by the media, Mayhew dismisses this quickly as the relationship between redistricting and electoral benefits for incumbents does not appear strong. The latter four hypotheses rely upon a central assumption: being an incumbent has, one way or another, become more electorally valuable than ever before for House members (Mayhew 2008).
This is not to say the value of incumbency is automatic, however. As Richard Fenno makes clear in his detailed case studies of House members’ activities in their districts: “[I]ncumbency is not an automatic entitlement to a fixed number of votes or percentage points towards reelection. Nor is ‘the power of incumbency’ something each member finds waiting to be picked up and put on like a new suit. Incumbency should be seen as a resource to be employed, an opportunity to exploited; and the power of the incumbency is whatever each member makes of the resource and the opportunity” (Fenno 2003, 211). Mayhew’s next three hypotheses are just that – describing how incumbents may have started making better use of their resources than previously leading to larger election margins.
His second hypothesis posits that incumbents have become better at gaining name recognition while in office with a particular focus on the uptick in franking between 1954 and 1970, shown in Figure 2 (Mayhew 2008). This hypothesis, at least when formulated around franking, quickly becomes unviable when it is brought up to speed with the current status of franking since “[r]eform efforts during the past 30 years have reduced overall franking expenditures in both election and non-election years . . . House mail costs have decreased from a high of $77.9 million in FY1988 to $15.1 million in FY2014” (Glassman 2015, 13). But this is not to rule out the name recognition hypothesis entirely as franking is only one of many ways an incumbent may get their name out more easily than a challenger.
The third hypothesis focuses on what Mayhew regards as a central activity of congressmen – credit claiming. Although there are various particularized government benefits available to be distributed by a congressman in their district, this hypothesis focuses on the rapid increase in federal grant-in-aid programs that appeared in the 1960s and 1970s. These grants, in theory, provide incumbents with an easy claim of benefitting part or all of their constituency. This growth in grants has by no means ceased since this hypothesis was written. Figure 3 shows total grants to state and local governments has continued to increase well after the initial explosion Mayhew observes (Mayhew 2008).
Fiorina contends that the reason district B ceased to be marginal so suddenly was due to the new Democrat’s tendency to focus less on national policy and more on district issues. He allocates more staff to district offices, meets with constituents more often than his predecessors, and does more constituent service. This latter trait is pointed to as the key to understanding the vanishing of the marginals. As size and scope of the federal government has expanded, so have opportunities for constituent service and the ability of congressmen to provide constituent service with larger staffs (Fiorina 1977). Fenno echoes this point: “For the member of Congress, it is a highly valued form of activity. Not only is constituent service universally recognized as an important part of the job in its own right. It is also universally recognized as powerful reelection medicine” (Fenno 2003, 101). This powerful reelection medicine could potentially have a side effect of vanishing marginals.
This expanding bureaucracy argument is also tied back more closely to Mayhew’s grant-in-aid hypothesis stating: “In addition to profitable casework let us remember too that the expansion of the federal role has also produced a larger pork barrel. The pork barreler need not limit himself to dams and post offices” (Fiorina 1977, 180). In Fiorina’s eyes, it is not just expanding grants-in-aid and credit-claiming that fuel the vanishing of the marginals but a broader expansion of the bureaucracy that both provides incumbents with more resources and with more constituent casework to win over their districts with (Fiorina 1977).
Mayhew’s fourth hypothesis focuses on another major activity of congressmen – position taking. Specifically, that congressmen have become much better at choosing positions which their district will agree with. His focus is on improvements in polling since the 1960s but this hypothesis could likely now be expanded to include the increased importance of data in politics more generally. However, this broadened explanation would prove more difficult to measure than the already difficult initial hypothesis.
Mayhew’s fifth and final hypothesis turns the attention from the behavior of the incumbent to the behavior of the voter. The claim is that dissatisfaction with party cues and a general downtick in partisanship has led voters to instead turn to nonpartisan cues instead – namely that of incumbency – and thus congressmen have not had to change their behavior in any way for the marginals to vanish (Mayhew 2008).
In their attempt to explain the vanishing marginals, Robert Goidel and Todd Shields expand this argument of growing dissatisfaction with party cues, pointing specifically to the advantage incumbents have in the media coverage. They posit that incumbents can, more easily than challengers, portray themselves as above partisanship and as servants of their district in the media. This, along with their ability to get more media coverage in general, increases the chances of an individual to vote for the incumbent when exposed to these nonpartisan cues in the media. Increased exposure to the media was found to correlate with increased probability of voting for the incumbent, most markedly among independents, using data from the 1990 NES study (Goidel and Shields 1994). This explanation also ties into Mayhew’s second hypothesis of name recognition, albeit in a way less under direct control of the incumbent than franking.
While Mayhew does not claim to offer a decisive explanation for the vanishing marginals, he provides a groundwork for investigating them. His work did not only spur searches for explanations of the vanishing marginals but also rebuttals questioning the very existence of them. One such rebuttal comes from Gary Jacobson: “[I]ncumbents are no safer now than they were in the 1950s, the marginals – properly defined – have not diminished, let alone vanished . . . competition for House seats held by incumbents has not, in fact, declined” (Jacobson 1987, 128).
On the point of incumbent safety Jacobson points out that, although for the 1960s “incumbent losses were abnormally high early in the decade and abnormally low at its end,” from decade to decade incumbents have not seen any increase in their likelihood of winning (Jacobson 1987, 128). The probability of winning, he argues, is more significant when determining electoral safety than electoral margins are – it does not matter how much a candidate wins by, only that they win. As such, although electoral margins grew wider, that does not mean fewer districts are prone to flip. When marginal elections are redefined as those elections in which an incumbent is likely to lose and not just those which have narrow margins the marginals no longer vanish (Jacobson 1987).
Although this assessment reframes things to demonstrate electoral competition has not declined, it does still leave a phenomenon to be explained: widening electoral margins. Jacobson believes this can be explained by the increase in the heterogeneity of interelection vote swings within districts. With decreasing reliance on party cues, there are fewer automatic votes for or against the incumbent based solely on their party, providing a bloc of voters who, previously voting their party, now rely on more candidate-centric cues like those discussed previously. In contrast to Mayhew’s fifth hypothesis, Jacobson does not make the case that turning away from party cues solely benefits the incumbent: “It allows members to develop a kind of personal hold on a district that helps insulate them from external political forces. But a less partisan, candidate-oriented electorate is also evidently more fickle, support from constituents is easier to lose as well as to win” (Jacobson 1987, 137).
In response to this rebuttal from Jacobson, Stephen Ansolabehere, David Brady, and Morris Fiorina counter in defense of vanishing marginals and declining electoral competition. They claim that the “Mayhew marginals” (districts where members win with 50-55% of the vote) did decline. The marginals that vanished from this category have since become “Jacobson marginals” – incumbents who, despite previous healthy margins, lost reelection and that this class of marginals has increased. They conclude: “[F]reshmen incumbents and/or those who have a close election typically redouble their efforts and improve their showings, while some long-time incumbents who achieve big victories grow overconfident, slacken their efforts and, consequently, do less well” (Ansolabehere, Brady, and Fiorina 1992, 27).
The result of this is the observation that marginal incumbents have gained some electoral safety while the safety of more comfortable incumbents has decreased slightly. In light of this, on balance, incumbents have not gained any electoral security as a result of the Mayhew marginals vanishing. What were previously considered safe incumbents are subject to “random terror” of being tossed out and as a result exhibit more responsiveness to their district (Ansolabehere, Brady, and Fiorina 1992).
Debate over the mechanisms behind, as well as the existence of, declining marginals and competition of House elections did not end in the 1990s. Alan Abramowitz et al. take particular note of the 2004 House elections which saw only 22 elections with a margin smaller than 10% with 172 candidates winning with 40% of greater margin or no opponent at all. From these statistics they conclude that “[t]he 2004 House elections were extraordinarily uncompetitive” (Abramowitz, Alexander, and Gunning 2006, 75).
Shying away from Jacobson’s redefinition of what constitutes a marginal election, they point to the precipitous decline in elections decided by a margin of 10% or less as a startling indication of decreased competitiveness; 22% of races from 1946 to 1950 met this definition of marginal with the rate steadily decreasing until just 7% of elections could be classified as marginal in the 2002-2004 elections. In an attempt to explain this, three explanations are proposed and investigated.
Their second potential explanation attributes the decline of marginals to increased partisan polarization. As polarization continues, red districts become redder and blue districts become bluer due to individuals increasingly moving to places consistent with their partisanship as well as immigration and individual ideological realignment. Their constructed hypothesis to test this explanations validity is: “If the partisan polarization hypothesis is correct, we should find that the number of safe districts has been steadily decreasing, with most of this change occurring between redistricting cycles” (Abramowitz, Alexander, and Gunning 2006, 77). In Figure 1, as the hypothesis predicts, the most significant changes in safe districts occurs between redistricting cycles.
Their third potential explanation is that of increasing incumbency advantage. They subdivide this explanation into two parts, advantages from holding office and advantages from incumbent campaigns, focusing primarily on the latter regarding campaign finances. This differentiates their explanation from Mayhew’s second and third hypotheses which looked at congressional franking and credit claiming from increases in federal grant-in-aid programs as mechanisms for increasing incumbency advantage – activities of the office rather than the campaign. The hypothesis they test is: “If the incumbency hypothesis is correct, we should find that even in marginal districts, competition is now relatively rare because of the inability to compete financially. We should also find that the decline in competition in House elections has been most evident in races involving incumbents, with competition declining less drastically in open seat races” (Abramowitz, Alexander, and Gunning 2006, 77).
They find that since 1952, while marginal races with an incumbent running declined, marginal open races saw no such decline. Further, they find that there is a significant correlation between how much a challenger to an incumbent spends and how likely they are to win. As the cost of running a campaign has increased, incumbents have kept up and put a greater focus on fundraising while challengers have been unable to match the fundraising of their counterparts in office.
Performing path analyses on the direct and indirect relationships of district partisanship (as a proxy for polarization), incumbency, and campaign spending on House election outcomes in six year time periods from 1972 to 2002 reveals: “Not only do a much larger proportion of House districts now strongly favor one party, but the majority party is more likely to be represented by an incumbent and to enjoy an overwhelming advantage in campaign spending. The result is a pattern of reinforcing advantages that leads to extraordinarily uncompetitive elections” (Abramowitz, Alexander, and Gunning 2006, 86). In short, they pin the decline in competition on the polarization of the electorate and financial advantages of incumbents.
This polarization, inherently tied to party loyalty, is a notable change from the America that Mayhew and others were examining in the 1970s and 1980s in which party loyalty seemed to be declining and thus the importance of party cues diminishing (Mayhew’s fifth hypothesis). As a result of the polarization of the 2000s, Jacobson presents evidence that the incumbency advantage is now declining in value, bucking the trend of the previous decades.
His data demonstrate a nationalization of House elections in recent years in which the linkage between presidential and congressional elections that had been waning reappeared as polarization took root. Districts in which an incumbent could win reelection despite their district being won by the opposite party in the presidential election have declined implying that whatever advantages inherent to an incumbent could pick them up votes from the opposite party are no longer doing so (Jacobson 2015).
This pattern is shown in Figure 5 in which reelection rates for incumbents in districts with presidential vote leaning to their party, the other party, or being balanced (within 2% of the national vote) are compiled by decade. While incumbents in districts leaning to their party have seen little change in their high reelection rates as expected, balanced and challenger party leaning districts both show an increase in reelection rates into the 1980s and then a decline afterwards consistent with the fall and then rise in party loyalty among voters.
Regarding the financial advantage of incumbents examined by Abramowitz et al., Jacobson contends that this too has eroded in recent years. Parties are now much more effective at doling out funds to any promising challengers to incumbents that appear. Beyond direct assistance from parties, Citizens United v. FEC opened the floodgates of funding for independent campaigns that can support challengers. These factors have granted challengers, at least in districts that don’t lean to the incumbent’s party, much better financial prospects.
These findings, Jacobson claims, provide considerable insight into not only the recent decline in incumbency advantage, but also for Mayhew’s vanishing marginals – they ultimately resulted from a decreased reliance on partisan cues. The other proposed incumbent activities examined here were a means to fill the void left by decreased party loyalty and allowed incumbents to gain votes previously out of reach. When party loyalty made a comeback as the country polarized, though many of these adaptive activities remained, they were trounced by party cues and the value of incumbency, although still present, declined (Jacobson 2015).
The vanishing marginals and their implications for electoral politics produced an impressive amount of work by scholars, the surface of which is barely scratched here. Even some four decades after the case of the vanishing marginals was first presented, it has not been decidedly cracked with multiple interacting factors likely at play. The efforts to find an explanation, however, have offered extensive insight into the driving forces behind House elections whether it be activities and styles of incumbents, broader changes in voter behavior (particularly polarization), or changes in the function of government.
References
Abramowitz, Alan I., Brad Alexander, and Matthew Gunning. 2006. “Incumbency, Redistricting, and the Decline of Competition in U.S. House Elections.” The Journal of Politics68(1): 75–88.
Ansolabehere, Stephen, David Brady, and Morris Fiorina. 1992. “The Vanishing Marginals and Electoral Responsiveness.” British Journal of Political Science22(01): 21.
Fenno, Richard F. 2003. Home Style: House Members in Their Districts (Longman Classics Series). Upper Saddle River: Pearson.
Fiorina, Morris P. 1977. “The Case of the Vanishing Marginals: The Bureaucracy Did It.” American Political Science Review71(01): 177–81.
Glassman, Matthew E. 2015. “Franking Privilege: Historical Development and Options for Change.”
Goidel, Robert K., and Todd G. Shields. 1994. “The Vanishing Marginals, the Bandwagon, and the Mass Media.” The Journal of Politics56(3): 802–10.
“Historical Tables.” The White House. https://www.whitehouse.gov/omb/historical-tables/ (May 10, 2019).
Jacobson, Gary C. 1987. “The Marginals Never Vanished: Incumbency and Competition in Elections to the U.S. House of Representatives, 1952-82.” American Journal of Political Science31(1): 126.
Jacobson, Gary C. 2015. “It’s Nothing Personal: The Decline of the Incumbency Advantage in US House Elections.” The Journal of Politics77(3): 861–73.
Mayhew, David Raymond. 2008. Parties and Policies: How the American Government Works. New Haven, CT: Yale University Press.
Comments