American democracy has long puzzled social scientists who have studied the topic at some length. On the one hand, the United States has maintained its democratic form for over 225 years and has, under most metrics, been quite successful under the sovereign indirect control of the electorate. On the other hand, political science research has consistently found the American voter to be surprisingly uninformed. Larry Bartels goes so far as to say that “the political ignorance of the American voter is one of the best documented data in political science” (1996, 194). But what follows from both observations seems to be quite conflicting. How is it that a democratic ship of state could ever function well, let alone even competently, when its captain (the American citizen) is so unqualified to steer it? Or in other words, how could democracy possible work under conditions of poor information and sophistication?

Democratic theorists and political scientists have pondered this tension and many have turned to retrospective voting as the effective mechanism which might assist the electorate. “Voters are not fools” argues Key (1966, 7) but just act differently from the metrics used to measure political sophistication and knowledge in the Michigan School of thought (Campbell, Converse, Miller, Stokes 1960). Instead of learning the policy platforms of different politicians and parties as well as to what degree these platforms are put into place once in power, voters are focused on results. Key argues “The impact of events from inauguration of an Administration to the onset of the next presidential campaign may affect far more voters than the fireworks of the campaign” (1966, 9).

While nicely elaborated upon by Key, retrospective voting is perhaps best understood by being put in contrast with prospective voting. Retrospective voting is a process by which voters consider the performance of those who held or hold political office and assesses if the outcomes of their actions while in office were satisfactory enough to re-elect the officer holder or not. In its purest form, the retrospective voter does not consider the available alternatives to the current political office holder when making a decision, but nearly exclusively on the performance of the office holder while in power. Prospective voting on the other hand, considers and attempts to anticipate future performance in office instead of past performance. Prospective voting incentivizes voters to consider multiple options for office as a means of choosing the best option according to the perceived future performance of each possible office holder.

Within the context of a limited sophistication among the American electorate, Anthony Downs was one of the first social scientists to adopt the retrospective model as the proper mechanism for mitigation between democratic governance and political knowledge (1957). Downs argued that not only was retrospective voting a good strategy for overcoming limited knowledge, but also that it was rational for democratic societies to function this way. For Downs, the costs associated with becoming informed are far too burdensome for the average citizen. Instead voters should look for ways to cut down on the costs of gaining information by using shortcuts including ideology and past performance. This approach is not a purely retrospective one, since Downs does suggest voters consider all options when voting, including those parties who have not held power and could not be retrospected upon. But he does suggest that retrospecting on the performance of the incumbent party remains one of the most accurate pieces of information available to the electorate and will be a vital part of the voter decision process.

While the concept of retrospective voting began to develop within the social science literature with Downs and Key, its value was implicitly suggested before either thinker with the 1950 APSA report “Toward a More Responsible Party System” (Committee on Political Parties 1950). Starting in the 1950’s throughout most of the 1970’s, political scientists worried political parties were nearing extinction within the American electorate and consequentially, its governance. Not only were parties perceived to have little control over its members to effectively govern, but the ideological foundations of both parties were murky and regularly encroached upon by members of the other party. One of the major reasons for the committees call for more ideologically distinct and homogeneous parties that acted upon their platform once in power was its ability to provide sufficient information for voters to engage in retrospective voting. Unlike a government of ad hoc coalitions passing laws in which the electorate cannot easily know who is responsible for the passage of any bill, with responsible parties the electorate knows who runs the state and can quickly judge the outcomes of their time in power at the voting booth. Retrospective voting was without a doubt a part of the discussion within the field of political science by the mid-19th century, but it did not receive much of a systematic empirical inquiry until Gerald Kramer’s article “Short-Term Fluctuations in U.S. Voting Behavior, 1896-1964,” in 1977. Kramer focused on retrospective voting towards incumbency oriented evaluations based off of the state of the economy while in office. Doing a time series analysis, Kramer found that electoral support for the incumbent party in Congress is best predicted by changes in the real per capita income measurement of the economy. But this incumbency variable was only important when the incumbent was of the same party as the President. Kramer’s work was one of many new publications at the time finding similar results correlating voting patterns and real per capita income (Bloom and Price, 1975; Tufte, 1975; Hibbing and Alford, 1981). A slightly different approach by Ray Fair found that both the unemployment rate as well as per capita GNP also worked nearly as well (1978). Meltzer and Vellrath also found a connection between the national unemployment rate and presidential vote totals, although it has been noted that the findings are significantly weaker than they had anticipated (1975). Different methodological approaches also contributed to the literature affirmatively correlating the state of the economy to quarterly presidential approval surveys. Study after study found that approval ratings for the incumbent President tended to be negatively affected by declines in either GNP or GDP output (Kernell, 1978; Golden and Poterba, 1980; Hibbs, Rivers, and Vasilators, 1982). There were obvious advantages to the use of presidential approval ratings over other measurable dependent variables which bolstered the retrospective voting theory, including the greater frequency of data collected compared to federal elections, which happen every two years for Congress and four years respectively for the President. The greater amount of data not only allowed for more robust analysis, but also allowed for more fine tuned measurements regarding how responsive the electorate was towards these economic indicators.

Because of the heavy emphasis on macro level data that social scientists had predominantly used for measuring retrospective voting, concerns began to arise that scholars were opening themselves up to the ecological fallacy. To combat this, further work on retrospective voting in the 1980’s considered the voters economic evaluations, not from the macro-level of GNP or the unemployment rate, but from the personal level of family finances. Morris Fiorina was one of the first to utilize survey data to see if a correlation existed between individual financial wellbeing and voting decisions, but his findings were disappointingly weak, showing a marginal correlation at best (1978). Roderick Kiewert also looked into the claim but similarly found little evidence correlating personal family finance to incumbency support (1983). Survey evidence also published that year added support to Fiorina’s and Kiewert’s findings showing that “virtually no one sees government policy contributing to their family’s economic achievements or setbacks. In any single national survey, no more than 1% of those interviewed point directly to government” (Kinder and Mebane 1983). Of course the survey evidence by itself may have been skewed by a desirability bias driven by American socio-cultural norms for individual responsibility, but coupled with Kiewert’s and Fiorina’s work it provided good evidence to rule out such personal level retrospection. Similar findings were found by other scholars throughout the decade, adding to the dissociation between personal finances and voting behavior (Feldman, 1982; Abramowitze et. al. 1988).

Following the weak findings between personal finances and retrospective voting, an interesting divide took place regarding the perceived overall value of the retrospective voting model as a theory for voting behavior. The difficulty remained that while retrospective voting seemed to take place at the macro level, it did not seem to have explanatory value at the micro level. The apparent lack of compatibility between these two findings led some scholars, including Gary Jacobson, to argue that what we were seeing was not the product of retrospective voting at all, but instead only a correlation between the overall state of the economy and the quality of challengers to incumbents (1989). Jacobson argued that when the economy is doing well, stronger candidates who will often already hold an office of power will not risk running against an incumbent. But when economic conditions begin to worsen, stronger candidates perceive the situation as opportune to challenge the incumbent. Because voters look at the strength of the candidate, higher rates of losing incumbents will result from poor economic conditions, not because voters took the economy into consideration, but because a stronger candidate challenged the incumbent.

Weighing in on the debate against retrospective voting, Brad Lockerbie suggests theoretical priority be given to the prospective voting model over the retrospective with his 1991 publication “Prospective Economic Voting in U.S. House Elections, 1956-88”. In his article, Lockerbie takes issue with some of the methodological variables utilized in the retrospective voting literature. Lockerbie argued that survey research measuring voter assessments of the economy near elections do not take into account a prospective look towards a new party taking power. If voters, who were holding the Republican party responsible for poor economic conditions anticipated a Democratic President would take over shortly, they might have a rosier view about the direction of the economy which might skew survey results about voter perceptions of the economy.

Despite the critique of scholars like Jacobson and Lockerbie, many researchers have continued to posit retrospective voting as the most appropriate interpretation of how voters vote. Kiewert himself, after being one of the main contributors to the empirical findings dissociating retrospective voting and personal finances, points towards national business conditions as having nearly twice the effect on incumbency support than family finances (Kinder, Kiewert, 1979; 1981). Kiewert and Kinder suggested that while voters may not connect their personal finances to the performance of their representatives while in power, voters might still perceive macro-level economic conditions and use these perceptions as a guide for voting. In response to Lockerbie’s criticisms of the literature, new research was also taken on with measures used to control for these variables mentioned by Lockerbie with an analysis re-done to measure for the effect of both prospective and retrospective voting (Lanoue 1994). The results showed that while prospective voting did appear to a greater extent then previously found, retrospective voting was still significantly more important.

Philip Jones recently added to the criticisms on retrospective voting with his 2011 study which used survey data to suggest that voters do not vote so much based off of the results of any policy, but on the policy itself and its connection to constituent ideology (Jones, 2011). Jones argues that most studies on retrospective models have failed to include a measure incorporating how well constituents have had their policy views represented by the incumbent. He proposes that when this variable is included, a much larger correlation exists between this variable and voting patterns than other outcome based factors of policy. What Jones fails to account for is how the electorate uses policy decisions of their representatives for guides to voting when the empirical evidence continues to show a lack of knowledge regarding the policy decisions of the electorate’s representatives. How does the electorate reward their representatives for votes that match their own ideological preferences if the electorate is unaware of how their representatives voted? Certainly an important question to answer before anything close to a definitive claim about causal factors at play in Jones’ findings correlating representative policy decisions with constituent ideology can be made. Furthermore, interesting new studies exist showing just how strong retrospective voting models can be, even when the deck seems stacked against the theory. The economy has long been found to be the most important variable in explaining voting behavior within the retrospective voting literature, implying that other variables should be of less importance or none at all. But what if variables that seem completely irrelevant to political performance were found to have correlation with voting behavior? This was exactly what a 2010 study discovered when exploring the relationship between college sporting events and voting behavior (Healy, Malhorta, Mo, 2010). The study found that college football teams that win within ten days of an election increase the votes of the incumbent representative of the district of that team by as much as 1.6%. Similarly, wins and loses during the NCAA college basketball tournament were found to be significantly correlated with presidential approval numbers. Implicitly within studies like this is a potential argument to be made that if such irrelevant factors as college sports influence the voting behaviors of the electorate, how much more so should we expect things like the economy to have influence over voting decisions?

Overall, the evidence considering retrospective voting remains compelling to many scholars studying voting behavior. In their own review of the retrospective voting literature in the mid 1980’s, Kiewiet and Rivers conclude that “a judicious summary of the empirical literature might run: A 1% decline in real income will cost the incumbent party between 0.5% and 1% of its vote share in the last election” (1984). Much of the research done on retrospective voting readily concedes that other variables, particularly party identity and incumbency tend to play a substantially larger role in influencing voting behavior (Lanoue, 1994); but this in no way negates the importance of retrospective voting since party identity and incumbency are relatively static variables. Retrospective voting based on other more frequently changing variables such as the economy might have explanatory power over short term changes in democratic outcomes.

One thing worth further consideration is the heavy emphasis on economics when studying retrospective voting over other areas of state control including foreign policy. This emphasis, as it exists in the literature, comes partly from the weak correlation in early research between foreign policy and retrospective voting and partly for reasons of convenience. First, early studies done between foreign policy affairs and voting decisions found weak relationships between the two which directed research towards economics because of the more promising finds (Stokes, 1966). And while later studies have slowly brought more credibility to the concept of foreign policy guided retrospective voting (Gelpi, Reifler, Feaver, 2007), it still remains much more muddled with prospective voting, making a nice demarcation between the two difficult to obtain. Also, studying foreign policy to understand retrospective voting has its disadvantages in comparison to studying economic factors which makes the latter more appeasing. Changing foreign policy dynamics make for many more unique variables within the data that becomes much more difficult to use and compare with different studies. We are not always at war or tied up in widely known foreign affairs. And when we are, the variables that make up these situations tend to look rather unique each and every time. Economic data, on the other hand remains relatively stable and easily measurable. Because economics have been using the same, or similar, measures to measure things like the unemployment rate of GDP, it has largely become a pragmatic choice among social scientists to measure economic variables over others if one wants to explore retrospective voting in general.

Considering retrospective voting in the abstract, and particularly as it relates to economic conditions, it remains a plausible explanation for successful democratic governance in the United States. But questions still loom for the retrospective model that need answering. First of all, although voters may to a certain degree vote retrospectively, do voters hold all elected officials accountable for these conditions or only certain offices? Secondly, what sort of time frame does a voter use for economic evaluation? Do they consider the state of the economy the day of the election when making their decisions or do they take into account a longer period of time? And finally, if retrospective voting is going to be a normative model for democratic societies, it should be asked how well can the average voter objectively retrospect on the state of the economy in the face of partisan and ideological bias? All three of these questions will need to be considered if we are going to accept the narrative that retrospective voting produces good governance. The first of these three questions brings us back to the APSA proposal to move towards more responsible parties. For retrospective voting is only effective if voters actually punish or reward those who are truly responsible for the outcomes under retrospection. Many theoretical examples could be imagined in which political actors such as the President are unfairly held responsible for actions performed by a Governor, Congress, or even their predecessor to the office. Holding the right people responsible is important, but it is also incredibly difficult to do in a governmental system of separated powers and an era of divided governance. The amount of information needed to know who should be held responsible for government is incredibly large and the costs demanding. And in congruence with what theory tells us should happen, empirical research does appear to suggest that the electorate does indeed have a difficult time attributing responsibility in American government. Cross-national analysis studies show that complicated institutional separations appear to blur political responsibility when matching policy to appropriate institutions (Anderson, 2000; Anderson, 2006). The theoretical implication of these findings being that we end up right back where we started, with voters not having sufficient knowledge to make retrospective voting decision, the very problem retrospective voting was meant to resolve in the first place. And yet, regardless of the theoretical difficulty this problem lays before us, much of the evidence still tells us that voters do indeed vote retrospectively. The question now becomes, how?

Edward Tufte was one of the first to systematically consider this point as the focus of his publication and found that changes in per capita income influenced the electoral results of incumbent congressional candidates during midterm elections (1975). Hibbing and Alford (1981) added to Tufte’s findings confirming the impact of retrospective voting upon congressional elections, but also finding that it was the incumbent members of Congress who were more at the mercy of economic fluctuations than those in the same party running for Congress that were not incumbents. While using aggregate data, evidence favorably leaned towards macro-level economic perceptions as driving retrospective voting at the Congressional level, some research showed that the President was held more accountable for family finances than other offices. Those who reported being financially well off were statistically more likely to vote for the incumbent party in presidential elections over those who reported being worse off or the same (Fiorina, 1978; Klorman, 1978; Kiewiet, 1983). But much less significant effects have been found in congressional elections (Fiorina, 1978; Kinder and Kieweit, 1981) or gubernatorial elections (Klorman, 1978). While methodological concerns exist for much of this research and its capacity to control for attitudinal variables, interesting relations regarding the uniqueness of the presidency remain. The idea that retrospective voting is a phenomenon that heavily or almost exclusively harms the incumbent President and his party makes sense intuitively. The President is by far the most visible elected official. Because of this, holding the President responsible demands fewer costs for the voter than it does to hold other, less visible office holders responsible. Despite some of the other evidence published by Tufte and Alford, other research published in the 1980’s found no significant correlation between multiple economic factors, including price inflation and real income changes and congressional district vote totals (Owens and Olsen, 1980). Even Alford and Hibbings, whose research showed that congressional elections were impacted by retrospective voting found that incumbents of the President’s party were more strongly affected by economic fluctuations than incumbents of the other party, even when the other party held control over one of the legislative houses (1981).

More recent research has brought more nuance to the question about retrospective voting and the Presidency, finding that other variables about those running for President have important causal impacts upon how and why voters vote the way they do. The connection between the President and his party, as well as the knowledge about the candidates running for President, seem to influence the degree of retrospective voting voters do. Nadeau and Lewis-Beck find that when candidates are popular or well-known, voters tend to vote retrospectively. But when candidates are new to the public spotlight, voters tend to disregard the economic performance of the candidates party previously and become prospective voters (2001). Campbell, Dettrey, and Yin find similar results in their research showing that open-seat elections for Presidency tend to be more prospective than when there is an incumbent (2010), although this depends on the perceptual closeness between the candidate and the outgoing President of the same party. Malhotra and Margalit find that retrospective voting appears to be highly contingent upon the expectations set up by those seeking office (2014). The higher the expectations, the greater the retrospection of the electorate will be once the representative goes up for re-election as an incumbent. But while all offices will set high expectations for themselves if elected to encourage the vote, higher offices (particularly the President) will be more visible in the expectations they set up on the campaign trail therefore making them more vulnerable to retrospective voting during their next time up for election.

More work is needed here to try and tie up the loose ends between retrospective voting, the Presidency, and other elected offices. But retrospective voting does appear to be more closely tied to the Presidency and his party, although contingent upon other factors. How these contingencies relate to other offices such as Congress remains a difficult question worthy of future research. But enough evidence does exist to make us pause. For as important as the President is to American politics and governance, the office is not the alpha and omega of American governance and any model of voting which acts as if it is may be misguided in its approach.

Within the limited extent that voters can and do try to hold office seekers responsible for economic conditions, another question comes to mind. What time frame do voters consider in an incumbent’s tenure of office for evaluating economic conditions? Do voters compare the state of the economy at the moment someone takes office to the moment of possible re-election to see if things are good or improved? This would demand voters both have a long memory as well as patience to wait until near the time of election to make up their minds. But if this is not the case then when are voters considering the economy for voting guidance? Some of the earliest studies on retrospective voting found that at the aggregate level economic conditions best predicted voting behavior in the second quarter of an election year (Fair, 1978). Stigler’s study found that if you attempt to increase the time frame for considering economic performance to two years (the term of a representative in Congress), the relationship between economic conditions and voting behavior weaken significantly (1973). The implications of these studies being that voters consider only the very recent past when making economic evaluations. On the other side of the equation, studies done between Gallup presidential surveys and economic conditions show that there is a lag between improving economic conditions and attribution of responsibility to those in power (Hibbs, 1982; Norpoth 1984). Voters looking to make economic evaluations appear to do so between the second and third quarters of the election year. A relatively small time frame considering the length of most political offices are between two and six years.

Furthermore, a recent study by Krause and Melusky find in their research that the more time previous office holders take away from politics, the less likely the electorate is to retrospectively vote on their economic record (2014). In this study, ex-Governors were studied to see how their economic records, when in office, impacted how they performed in elections for different offices. They found that when ex-governors led the state during good economic times, they would receive about 68% of the vote share when running for a different office within two years of leaving the governorship. But the more time ex-Governors take between holding office and running for a new office, the less successful they were in their new endeavors. Just 45.5% of the vote went to ex-Governors with a strong economic record once four to twelve years had passed between holding the governorship and their running for a new office. If these studies are correct, then not only do voters seem to consider economic conditions within a very small window of the office holder’s tenure, but these evaluations of an office holder’s ability in office quickly lose importance after a few years out of office.

The literature on retrospective voting has brought to light a couple of trends to ponder. One being how the electorate assigns responsibility to the correct officeholders and the other being when evaluations are made for assigning responsibility. On both questions, research produces a picture for us that should make us question just how well retrospective voting can work as a model for democracy. But one more question needs to be considered. For nowhere in the literature evaluated has partisanship been replaced as a prominent causal factor for voting behavior. In fact, in Lanoue’s study of retrospective voting, which included a measure for partisanship, Lanoue found partisanship to be the strongest indicator of voting behavior, far above that of retrospective voting (1994). But for the room made for retrospective voting in voting behavior as a contributing factor, we are left to wonder, how much influence do variables like partisanship have upon our ability to retrospectively vote?

At the very least, we have known since the publication of the American Voter that the electorates identification with a political party heavily influences the way in which they vote (Campbell, Converse, Miller, Stokes 1960). Some recent research has even suggested that our party identification might be the largest causal contributor to our ideologies (Goren, 2005). Quite a fascinating find that implies the powerful influences party identification may have on how we construct and make sense of the world around us. But this should leave us wondering, how might party identity causally influence the way we interpret economic conditions when engaging in retrospective voting? One of the first to take up this question between party identification and retrospective voting was Fiorina, who suggested that party identification doesn’t influence our retrospective voting, but instead the exact opposite is happening (1981). Fiorina posited that voters retrospect on the performance of parties and keep a “running tally” between the two parties when identifying oneself with a party. So long as one party holds a higher score in the retrospective tally, a person will identify themselves with that party. Fiorina’s research has come under scrutiny with the publication of Eric Groenedyk’s article which provides evidence to the contrary, that party identification does impact the way we retrospectively vote (2011). Groenedyk’ s research showed that when subjects who identify with a party are given information about a candidate of their party that is inconsistent with the subject’s views, subjects have a tendency to lower the value of the opposition party further to maintain their party identification. Groenededyk calls this “the lesser of two evils” approach by which they actively restructure their understanding of the world around them to maintain their party identification.

While never specifically touched upon by Groenedyk, his findings have some major implications for the effectiveness of retrospective voting. If party identification is indeed an important part of how the electorate interprets information, then the prospect of voters objectively voting based off of economic conditions is significantly diminished. Voters utilizing the “lesser of two evils” method may recognize the poor state of the economy under the control of their party, but may not hold incumbents responsible because they arbitrarily construct challengers of the other party as still being worse. Here voters would think they are voting based off of the performance of current office holders by hypothetically considering how the other party would have done if they were in charge and arbitrarily concluding they would have done worse. But in reality voters would be doing nothing more than voting based off of party identity. A similar critique might be found elsewhere in Jones’ study, which found an interesting correlation between party identification and perceptions of the economy and representative congruency with personal ideology (2011). Jones found that for those represented by Democrats, the better they believe the economy to have gotten, the less likely they were to have believed they were being well-represented. But the exact opposite was found for those being represented by Republicans, for the better they believed the economy to be, the more likely they were to believe they were being well-represented. Why the party in office seems to have a correlative effect upon how voters perceive either the state of the economy or how well they are being represented is difficult to understand. But the possible implications may be a lack of capacity for voters to remove their partisanship bias and objectively consider conditions worthy of retrospection when voting.

So what should we make of retrospective voting as a proposed model to explain the success of democratic governance? We first explored if retrospective voting was actually a method utilized by the American voter and to a certain level of confidence find that the literature does show it has explanatory value for voting behavior. We then considered if it could be used as a model to explain “good voting”, or voting patterns that explain the success of American democracy. Three stumbling blocks appear in the literature that diminish one’s hope for retrospective voting. The first being the difficultly of assigning rewards and blame in the American system of federalism and the separation of powers. It is uncertain how voters decide who to hold responsible, although compelling evidence exists that it is probably President focused. The second stumbling block was figuring out when exactly voters are making evaluations of economic performance? Despite the two to six year tenure most political offices have, the literature provides compelling evidence that evaluations are made between the second and third quarters of an election year. This means the performance of office holders before this time frame does not seem to hold much weight, nor what they accomplish between the third quarter and election day. The final stumbling block retrospective voting needs to overcome is the capacity for voters to objectively evaluate economic conditions without the heavy influence of party identification. If voters are heavily influenced by their party identification when considering the economic performance of elected officials, then what we will most likely see is not retrospective voting at all, but instead partisan voting that is only ad hoc justified in the minds of the voter as retrospective.

A few research questions come to mind based off of the literature explored above. The first is how much does party identification influence voter perceptions about the state of the economy? This research question would explore the idea of cognitive bias and its contribution through party identification as a contributor to economic evaluation. Groenedyk’s publication explores cognitive bias and how we diminish the valuations of the opposite party to maintain party identification, but he never explores how party identification influences our valuations of other things other than the parties themselves. Should we expect that someone who identifies with the Democratic party will evaluate the economy better if the Democratic party runs government than if the Republican party were running government with the same economic indicators, and vice-versa? If so, how would they do so? Exploring this question in research would be difficult to do. While utilizing survey data would have many benefits, we currently live in an era of divided government, which means regardless of one’s perception of the state of the economy, it would be difficult to parse out if they contribute economic conditions to one party or the other. Also, we want to explore if participants would look at the same economic situation the same of differently depending on their partisanship. This would be difficult to do with survey data since we cannot know what different individuals will consider when thinking about the economy. Will they take into account personal finances? National GDP? Income inequality? We should expect different partisan affiliations to direct voters to use different economic factors in their evaluations of economic conditions, but this would not be the point of the study. What I would want to see is how voters consider the same economic variables when making judgements about the economy. Instead, laboratory experiments would probably need to be used where participants might be given hypothetical situations that include economic data given for both the beginning and end of a parties term in office. Participants would probably be asked to give an evaluation on how well the party did or did not do while in office and match this to pre-tests about the participants party identity. I would probably utilize close ended and open ended questions as well as questions that utilize a 1 to 10 scale for measurements. This would allow us to give the same information about the state of the economy for all to see what differences might come about in evaluation because of partisanship.

The second research question that arises from my literature review would be, do office holders believe the electorate utilizes retrospective voting or not? That is, do office holders believe that the state of the economy influences their chances at re-election? To measure this, I would look to analyze previous incumbent campaigns to see to what extent incumbents mention economic conditions in their campaigns both when the economy is doing well and when it is doing poorly. If incumbents believe that voters vote retrospectively, we should expect incumbents to mention economic conditions significantly more when the economy is doing well and significantly less when the economy is doing poorly. On the other hand, challengers should be expected to mention economic conditions much more when the economy is doing poor and much less when the economy is doing well. Campaign speeches, slogans, ads, and other campaign material would need to be analyzed and coded for to run some sort of statistical regression on our findings.


Abramowitz, Alan I., David J. Lanoue, and Subha Ramesh. (1988). ‘‘Economic Conditions, Causal Attributions, and Political Evaluations in the 1984 Presidential Election.’’ Journal of Politics 50: 848–63. Anderson, Christopher JH. (2000). “Economic Voting and Political Context: A Comparative Perspective.” Electoral Studies 19: 151 - 170. Anderson, Cameron D. (2006). “Economic Voting and Multilevel Governance: A Comparative Individual-Level Analysis.” American Journal of Political Science 50:449-463. Bartels, Larry M. 1996. “Uninformed Votes: Information Effects in Presidential Elections.” American Journal of Political Science 40: 194-230. Bloom, Howard S., and H. Douglas Price. (1975). “Voter Response to Short-run Economic Conditions: The Asymmetric Effect of Prosperity and Recession.” American Political Science Review 69: 1240 -1254. Campbell, A., Converse, P.E., Miller, W.E., & Stokes, D.E. (1960). The American Voter. New York: Wiley. Campbell, James E., Bryan J. Dettrey., and Hongxing Yin. (2010). “The Theory of Conditional Retrospective Voting: Does the Presidential Record Matter Less in Open-Seat Elections?” The Journal of Politics 72: 1083- 1095. Committee on Political Parties, American Political Science Association. (1950). “Toward a More Responsible Two-Party System”. New York: Rinehard. Downs, Anthony. An Economic Theory of Democracy. (1957). Boston: Addison-Wesley Publishing. Fair, Ray C. (1978). “The Effect of Economic Events on Votes for President.” The Review of Economics and Statistics 60: 159-173. Feldman, Stanley. (1982). “Economic Self-Interested and Political Behavior.” American Journal of Political Science 26:446 - 466. Fiorina, Morris P. (1978). “Economic Retrospective Voting in American National Elections: A Mirco-Analysis.” American Journal of Political Science 22: 426 - 443. Fiorina, Morris P. (1981). Retrospective Voting in American National Elections. New Haven: Yale University Press. Gelphi, Christopher., Jason Reifler and Peter Feaver. (2007). “Iraq the Vote: Retrospective and Prospective Foreign Policy Judgments on Candidate Choice and Casualty Tolerance.” Political Behavior 29:151-174. Golden, D.G. and Poterba, J.M. (1980). “The Price of Popularity: The Political Business Cycle Reexamined.” American Journal of Political Science 24: 696-714. Goren, Paul. (2005). “Party Identification and Core Political Values.” American Journal of Political Science 49: 881-896. Groenendyk, Eric. (2011). “Justifying Party Identification: A Case of Identifying with the “Lesser of Two Evils”.” Political Behavior 34:453-475. Healya, Andrew J., Neil Malhotrab., and Cecilia Mo. (2010). “Irrelevant Events Affect Voters Evaluations of Government Performance.” Proceedings of the National Academy of Sciences 107: 12804- 12809. Hibbing, John R., and John R. Alford. (1981). “The Electoral Impact of Economic Conditions: Who is Held Responsible?” American Political Science Review 25: 423-439. Hibbs, Douglas A., Douglas Rivers, and Nicholas Vasilatos. (1982). “On the Demand for Economic Outcomes: Macroeconomic Performance and Mass Political Support in the United States, Great Britian, and Germany.” Journal of Politics 43: 426-462. Jacobson, Gary. (1989). “Strategic Politicians and the Dynamics of U.S. House Elections, 1946-86.” American Political Science Review 83:773-793. Jones, Philip E. (2011). “Which Buck Stops Here? Accountability for Policy Positions and Policy Outcomes in Congress.” The Journal of Politics 73: 764-782. Kramer, Gerald H. (1977). “Short-Term Fluctuations in U.S. Voting Behavior, 1896-1964.” American Political Science Review 65: 131-143. Kernell, Sam. (1978). “Explaining Presidential Popularity.” American Political Science Review 72:506 - 522. Key, V. O. 1966. The Responsible Electorate. Cambridge: Harvard University Press. Kiewert, D Roderick. (1983). Macroeconomics and Micropolitics. Chicago: University of Chicago Press. Kiewiet, Roderick., Douglas Rivers. (1984). “A Retrospective on Retrospective Voting.” Political Behavior 6: 369- 393. Kinder, Donald R. , and d. Roderick Kiewiet. (1979). “Economic Discontent and Political Behavior: The Role of Personal Grievances and Collective Economic Judgements in Congressional Voting.” American Journal of Political Science 23: 495- 517. Kinder, Donald R. , and d. Roderick Kiewiet. (1981). “Sociotropic Politics: The American Case.” British Journal of Political Science 26: 446- 466. Kinder, Donald R., and Walter Mebane. (1983). “Politics and Economics in Everyday Life.” In Kristen Monroe (ed.), The Political process and Economic Change. New York: Agathon Press. Klorman, Ricardo. (1978). “Trend in Personal Finances and the Vote.” Public Opinion Quarterly 42: 31-48. Krause, George A., and Benjamin F. Melusky. (2014). “Retrospective Economic Voting and the Intertemporal Dynamics of Electoral Accountability in the American States.” The Journal of Politics 76: 1102 - 1115. Lanoue, David J. (1994). “Retrospective and Prospective Voting in Presidential-Year Elections.” Political Science Quarterly 47:193 - 205. Lockerbie, Brad. (1991). “Prospective Economic Voting in U.S. House Elections, 1956-88.” Legislative Studies Quarterly 16: 239-261. Malhorta, Neil., and Yotam Margalit. (2014). “Expectation Setting and Retrospective Voting.” The Journal of Politics 76: 1000- 1016. Meltzer, Allan H., and Marc Vellrath. (1975). “The Effects of Economic Policies on Votes for the Presidency: Some Evidence for Recent Elections.” Journal of Law and Economics 18:781-798. Murakami, Michael. (2008). “Paradoxes of Democratic Accountability: Polarized Parties, Hard Decisions, and No Despot to Veto.” Critical Review 20: 91 - 113. Nadeau, Richard., and Michael S. Lewis-Beck. (2001). “ National Economic Voting in the U.S. Presidential Elections.” Journal of Politics 63: 159-181. Norpoth, Helmut. (1984). “Economics, Politics, and the Cycle of Presidential Popularity.” Political Behavior 6: 253- 274. Owens, John R., and Edward C. Olson. (1980). “Economic Fluctuations and Congressional Elections.” American Journal of Political Science 24: 469- 493. Stigler, George J. (1973). “General Economic Conditions and National Elections.” American Economic Review 63: 160-167. Stokes, D.E. (1966). “Some Dynamic Elements of Contests for the Presidency” American Political Science Review 60: 19-28. Tufte, Edward R. (1975). “Determinants of the Outcomes of Midterm Congressional Elections.” American Political Science Review 69: 812 - 826.