Psephology and forecasting election results based on opinion polls is well established. But is there a data driven election analysis? The question came to mind as Goldman Sachs came out with a report on Indian Elections 2014 and the Congress Party central government minister were very angry and said why stock broking firm is making election forecasts. Can a security analyst use methods similar to the ones he uses to make stock market target price predictions and price trends. Such methods will be called data-based methods.
An interesting blog post in the regard is Using Data-Driven Models to Predict Election Results
Fundamental Analysis and Models for Forecasting Elections
Fundamental models for forecasting elections are models that can make forecasts of the results of elections using only economic and political data available months before the election.
These models provide accurate forecasts of the results of elections before polls on voting intentions can accurately forecast elections. Fundamental models are also useful in that they give a better sense of what factors are driving the outcomes behind elections by indicating which types of economic and political data most meaningfully correlate with election outcomes.
Literature on Forecasting Elections
Arrow et al. (2008) notes that prediction markets resulted in an average forecasted error of 1.5 percentage points in the national popular vote in recent U.S. presidential elections, while the final Gallup poll resulted in an average error of 2.1 percentage points, and Berg et al. (2008a) finds similar results on the average accuracy of prediction markets and polls for forecasting the national popular vote just before U.S. presidential elections.
Holbrook and DeSart (1999), Kaplan and Barnett (2003), and Soumbatiants et al. (2006) show that one can use polls taken just before an election to make fairly accurate forecasts of the state-level results in U.S. presidential elections, and Rothschild (2009) shows this for both polls and prediction markets.
While polls and prediction markets can both be used to make reasonably accurate forecasts
of election results just before an election, these methods are much less reliable when used months
before an election.
Gelman and King (1993) notes that polls for the national popular vote in U.S. presidential elections tend to oscillate wildly in the months before the election takes place, and as a result, polls can be highly unreliable indicators of the election outcomes months before the election.
Arrow et al. (2008) further notes that using prediction market data to make forecasts of the national
popular vote five months before a U.S. presidential election would have resulted in an average error
of over 5 percentage points in recent elections, and that using pre-election polls would have resulted
in even less accurate predictions.
Rothschild (2009) investigates the errors in both polls and prediction markets at forecasting probabilities of victory at the state level up to 130 days before the election, and notes that there is not enough liquidity to even have predictions for some states.
Researchers have investigated many unique methods of forecasting election results months
before the election that may hold more promise than using polls or prediction markets. Several
techniques have been explored for forecasting the results of elections such as using biographical
See, for example, Arrow et al. 2008, Berg et al. 2008a; 2008b, Brown and Chappell 1999, Holbrook and DeSart 1999, Kaplan and Barnett 2003, Pickup and Johnston 2008, and Soumbatiants et al. 2006. Erikson and Wlezien (2008a) indicates that the inaccuracy of polls suggests a need to systematically adjust for biases
information (Armstrong and Graefe 2011), using measures of how well candidates would be
expected to handle particular issues (Graefe and Armstrong 2012), surveying experts or voters for
their predictions (Jones et al. 2007; Lewis-Beck and Tien 1999), using indices that reflect a variety
concerns such as whether there has been a major policy change, a major military failure or success,
and social unrest or a scandal (Armstrong and Cuzán 2006; Lichtman 2008), or even using pictures
or silent video clips of candidates (Armstrong et al. 2010; Benjamin and Shapiro 2009).
However, while there is an extensive literature on forecasting elections using econometric models, so far the vast majority of this literature has focused on forecasting nationwide results. This holds for forecasting models of presidential elections, which typically focus on forecasting the national popular vote, and for forecasting models of congressional elections, which typically focus on forecasting the number of seats won by each major party in the two branches of Congress.
The focus on forecasting nationwide results for these elections is somewhat unsatisfying
because the results of these elections are typically determined at the state or local level. For
instance, the Electoral College elects the U.S. president, where each state has electors that equal its
congressional representation. Further, gubernatorial and senatorial elections are also state-level
elections. Despite this fact, so far only a handful of papers have addressed questions related to
forecasting election outcomes at the state level using econometric methods. Several papers related
to forecasting the results of the U.S. presidential election at the state level are of limited practical
use for forecasting elections because they focus on showing theoretically how one might make
See, for example, Abramowitz 2008, Alesina et al. 1996, Bartels and Zaller 2001, Campbell 2008, Cuzán and Bundrick 2008, Erikson and Wlezien 2008, Fair 2009, Haynes and Stone 2004, Hibbs 2008, Holbrook 2008, Lewis-Beck and Tien 2008, Lockerbie 2008, Norpoth 2008, and Sidman et al. 2008.
5 See, for example, Abramowitz 2010, Abramowitz and Segal 1986, Bafumi et al. 2010a, Campbell 2010, Coleman 1997, Cuzán 2010, Fair 2009, Kastellec et al. 2008, Lewis-Beck and Rice 1984; 1985, Lewis-Beck and Tien 2010, and Marra and Ostrom 1989
forecasts of elections if certain data that is only available after elections were available before the
election (Rosenstone 1983; Holbrook 1991; Strumpf and Phillipe 1999).
Campbell (1992) and Campbell et al. (2006) illustrate how one can combine the results of
polls taken roughly two months before the election with a variety of other economic and political
indicators to make forecasts of the results of U.S. presidential elections at the state level.
Like Campbell (1992) and Campbell et al. (2006), Klarner (2008) uses information from pre-election polls on voting intentions in his forecasting model, though Klarner (2008) uses polls on voting intentions that are taken further in advance of the election than those considered by Campbell (1992) and Campbell et al. (2006). Klarner (2008) does not report the results of the errors in his out-of-sample forecasts in his paper.
Bardwell and Lewis-Beck (2004) report results of a model that forecasts the results of Senate elections in Maine, and Klarner (2008) develops a model that forecasts the results of Senate elections for all states using information from pre-election polls on voting intentions.
There have been papers that have investigated factors that affect vote choice in gubernatorial elections (Adams and Kenney 1989; Atkeson and Partin 1995; Carsey and Wright 1998; Hansen 1999; Howell and Vanderleeuw 1990; Niemi et al. 1995; Partin 1995; Peltzman 1987; Svoboda 1995),
Some Important Variables
Incumbency: In the case of presidential elections, we expect voters to react differently depending on which party is the incumbent president. For this reason, we include a dummy variable that equals 1 (-1) if a Democrat (Republican) is president. However, since there is empirical evidence that voters are less likely to want to reelect members of the incumbent party if the incumbent party has been in office for multiple terms (Abramowitz 2008), we also include a variable that equals 1 (-1) if the Democrats (Republicans) have been in control of the presidency for at least eight consecutive years and 0 otherwise.
There is also empirical evidence that voters vote differently in Senate elections that takes place on a midterm than they do in Senate elections that take place the same year as a presidential election. Busch (1999), Chappell and Suzuki (1993) and Grofman et al. (1998) all suggest that voters are less likely to vote for members of the president’s party in Senate elections during a midterm than they are during a year when there is a presidential election.
Past Election Results: One of the best indicators of how states will vote in the future is differences
in how states voted in previous elections. Thus, in forecasting the Electoral College, we consider
variables that represent the difference between the fraction of the major party vote received by the
Democratic candidate in the state and the fraction of the major party vote received by the
Democratic candidate nationwide, in both of the two previous presidential elections.
While these variables are helpful in forecasting the results of future elections, these variables
can also sometimes give a misleading picture of the ideologies of the states. In some previous
presidential elections, there was a major third party candidate who took substantially more votes
from one major party candidate than another. For this reason, we include variables that address how
the vote shares of major third party candidates in previous elections should affect the forecasts we
make for future presidential elections. In particular, we consider the three different third party
candidates who received more than five percent of the national popular vote when they ran for
State Ideology: In addition to past election results, we also find it helpful to include other measures
Different dummy variables for the different types of job experience that a candidate may have had has not been used before in the literature. The one other model for forecasting Senate elections at the state level that includes biographical information about the candidates instead considers a single variable that may assume
any one of several different arbitrarily chosen values depending on the previous experience of the
candidates (Klarner 2008).
While past election results are significant and meaningful in the presidential and senatorial
forecasting models, they are not statistically significant in the gubernatorial model. Including a
variable for past presidential elections analogous to that considered in the presidential and senatorial
models is not statistically significant in the gubernatorial model, perhaps a reflection of the fact that
gubernatorial elections and presidential elections involve different issues and voting patterns in one
of these types of elections are not especially predictive of voting patterns in the other. In addition,
The changes in income are better predictors of election results than gauges of the absolute level of performance of the economy such as, for example, absolute levels of
Abramowitz, Alan I. 1988. “Explaining Senate Election Outcomes.” American Political Science
Review 82, 2, 385-403.
Abramowitz, Alan I. 2008. “Forecasting the 2008 Presidential Election with the Time-for-Change
Model.” P.S. Political Science and Politics 41, 4, 691-695.
Abramowitz, Alan I. 2010. “How Large a Wave? Using the Generic Ballot to Forecast the 2010
Midterm Elections.” P.S. Political Science and Politics 43, 4, 631-632.
Abramowitz, Alan I. and Jeffrey A. Segal. 1986. “Determinants of the Outcomes of U.S. Senate
Elections.” Journal of Politics 48, 2, 433-439.
Abrams, Burton A. and James L. Butkiewicz. 1995. “The Influence of State-Level Economic
Conditions on the 1992 U.S. Presidential Election.” Public Choice 85, 1, 1-10.
Adams, James D. and Lawrence W. Kenny. 1989. “The Retention of State Governors.” Public
Choice 62, 1, 1-13.
Alesina, Alberto, John Londregan, and Howard Rosenthal. 1996. “The 1992, 1994, and 1996
Elections: A Comment and a Forecast.” Public Choice 88, 1, 115-125.
Alvarev, Michael R. and Jonathan Nagler. 1995. “Economics, Issues, and the Perot Candidacy:
Voter Choice in the 1992 Presidential Election.” American Journal of Political Science 39, 3, 714-
Ansolabehere, Stephen and James M. Snyder Jr. 2002. “The Incumbency Advantage in U.S.
Elections: An Analysis of State and Federal Offices, 1942-2000.” Election Law Journal 1, 315-339.
Armstrong, J. Scott and Alfred G. Cuzán. 2006. “Index Methods for Forecasting: An Application to
the American Presidential Elections.” Foresight: The International Journal of Applied Forecasting
Armstrong, J. Scott and Andreas Graefe. 2011. “Predicting Elections from Biographical Information
About Candidates: A Test of the Index Method.” Journal of Business Research 64, 7, 699-706.
Armstrong, J. Scott, Kesten C. Green, Randall J. Jones, and Malcolm Wright. 2010. “Predicting
Elections from Politicians’ Faces.” International Journal of Public Opinion Research 22, 511-522.
Arrow, Kenneth J., Robert Forsythe, Michael Gorham, Robert Hahn, Robin Hanson, John O.
Ledyard, Saul Levmore, Robert Litan, Paul Milgrom, Forrest D. Nelson, George R. Neumann,
Marco Ottaviani, Thomas C. Shelling, Robert J. Shiller, Vernon L. Smith, Erik Snowberg, Cass R.
Sunstein, Paul C. Tetlock, Philip E. Tetlock, Hal R. Varian, Justin Wolfers, and Eric Zitzewitz.
2008. “The Promise of Prediction Markets.” Science 320, 5878, 877-878.
Atkeson, Lonna Rae and Randall W. Partin. 1995. “Economic and Referendum Voting: A
Comparison of Gubernatorial and Senatorial Elections.” American Political Science Review 89, 1,
Bafumi, Joseph, Robert S. Erikson, and Christopher Wlezien. 2010a. “Forecasting House Seats
from Generic Congressional Polls: The 2010 Midterm Election.” P.S. Political Science and Politics
43, 4, 633-636.
Bafumi, Joseph, Robert S. Erikson, and Christopher Wlezien. 2010b. “Ideological Balancing,
Generic Polls and Midterm Congressional Elections.” Journal of Politics 72, 3, 705-719.
Bardwell, Kedron and Michael S. Lewis-Beck. 2004. “State-Level Forecasts of U.S. Senate
Elections.” P.S. Political Science and Politics 37, 4, 821-826.
Bartels, Larry M. and John Zaller. 2001. “Presidential Vote Models: A Recount” P.S. Political
Science and Politics 34, 1, 9-20.
Benjamin, Daniel J. and Jesse M. Shapiro. 2009. “Thin-Slice Forecasts of Gubernatorial Elections.”
Review of Economics and Statistics 91, 3, 523-536.
Berg, Joyce, Robert Forsythe, Forrest Nelson, and Thomas Rietz. 2008a. “Results from a Dozen
Years of Elections Futures Markets Research.” In Handbook of Experimental Economics Results,
Volume 1, eds. Charles R. Plott and Vernon L. Smith. North-Holland: Amsterdam.
Berg, Joyce A., Forrest D. Nelson, and Thomas E. Reitz. 2008b. “Prediction Market Accuracy in
the Long Run.” International Journal of Forecasting 24, 2, 285-300.
Brown, Lloyd B. and Henry W. Chappell Jr. 1999. “Forecasting Presidential Elections Using
History and Polls.” International Journal of Forecasting 15, 2, 127-135.
Bullock, Charles S. III. 1988. “Regional Realignment from an Officeholding Perspective.” Journal
of Politics 50, 3, 553-574.
Busch, Andrew E. 1999. Horses in Midstream: U.S. Midterm Elections and Their Consequences,
1894-1998. University of Pittsburgh Press: Pittsburgh.
Campbell, James E. 1992. “Forecasting the Presidential Vote in the States.” American Journal of
Political Science 36, 2, 386-407.
Campbell, James E. 2008. “The Trial-Heat Forecast of the 2008 Presidential Vote: Performance and
Value Considerations in an Open-Seat Election.” P.S. Political Science and Politics 41, 4, 697-701.
Campbell, James E. 2010. “The Seats in Trouble Forecast of the 2010 Elections to the U.S. House.”
P.S. Political Science and Politics 43, 4, 627-630.
Campbell, James E., Syed Ali, and Farida Jalalzai. 2006. “Forecasting the Presidential Vote in the
States, 1948-2004: An Update, Revision, and Extension of a State-Level Presidential Forecasting
Model.” Journal of Political Marketing 5, 1, 33-57.
Campbell, James E. and Joe A. Summers. 1990. “Presidential Coattails in Senate Elections.”
American Political Science Review 84, 2, 513-524.
Canes-Wrone, Brandice, David W. Brady, and John F. Cogan. 2002. “Out of Step, Out of Office:
Electoral Accountability and House Members’ Voting.” American Political Science Review 96, 1,
Carsey, Thomas M. and Gerald C. Wright. 1998. “State and National Factors in Gubernatorial and
Senatorial Elections.” American Journal of Political Science 42, 3, 994-1002.
Chappell, Henry W. Jr. and Motoshi Suzuki. 1993. “Aggregate Vote Functions for the U.S.
Presidency, Senate and House.” Journal of Politics 55, 1, 207-217.
Coleman, John J. 1997. “The Importance of Being Republican: Forecasting Party Fortunes in House
Midterm Elections.” Journal of Politics 59, 2, 497-519.
Cuzán, Alfred G and Charles M. Bundrick. 2008. “Forecasting the 2008 Presidential Election: A
Challenge for the Fiscal Model.” P.S. Political Science and Politics 41, 4, 717-722.
Cuzán, Alfred G. 2010. “Will the Republicans Retake the House in 2010?” P.S. Political Science
and Politics 43, 4, 639-641.
Erikson, Robert S. 1988. “The Puzzle of Midterm Loss.” Journal of Politics 50, 4, 1011-1029.
Erikson, Robert S. and Christopher Wlezien. 2008a. “Are Political Markets Really Superior to Polls
as Election Predictors?” Public Opinion Quarterly 72, 2, 190-215.
Erikson, Robert S. and Christopher Wlezien. 2008b. “Leading Economic Indicators, the Polls, and
the Presidential Vote.” P.S. Political Science and Politics 41, 4, 703-707.
Fair, Ray C. 2009. “Presidential and Congressional Vote-Share Equations.” American Journal of
Political Science 53, 1, 55-72.
Fiorina, Morris. 2003. Divided Government. Pearson Longman: New York.
Garand, James C. 1988. “Localism and Regionalism in Presidential Elections.” Western Political
Quarterly 41, 1, 85-103.
Gelman, Andrew and Gary King. 1993. “Why Are American Presidential Election Campaign Polls
So Variable When Votes Are So Predictable?” British Journal of Political Science 23, 4, 409-451.
Graefe, Andreas and J. Scott Armstrong. 2012. “Predicting Elections from the Most Important
Issue: A Test of the Take-the-Best Heuristic.” Journal of Behavioral Decision Making 25, 1, 41-48.
Grofman, Bernard, Thomas L. Burnell, and William Koetzle. 1998. “Why Gain in the Senate but
Midterm Loss in the House? Evidence from a Natural Experiment.” Legislative Studies Quarterly
23, 1, 79-89.
Hansen, Susan B. 1999. ““Life Is Not Fair”: Governors’ Job Performance Ratings and State
Economies.” Political Research Quarterly 52, 1, 167-188.
Haynes, Stephen E. and Joe A. Stone. 2004. “‘Guns and Butter’ in U.S. Presidential Elections.”
Economics Bulletin 1, 5.
Hibbs, Douglas A. Jr. 2008. “Implications for the ‘Bread and Peace’ Model for the 2008 US
Presidential Election.” Public Choice 137, 1, 1-10.
Highton, Benjamin. 2000. “Senate Elections in the United States, 1920-94.” British Journal of
Political Science 30, 3, 483-506.
Holbrook, Thomas M. 1991. “Presidential Elections in Space and Time.” American Journal of
Political Science 35, 1, 91-109.
Holbrook, Thomas M. 2008. “Incumbency, National Conditions, and the 2008 Presidential
Election.” P.S. Political Science and Politics 41, 4, 709-712.
Holbrook, Thomas M. and Jay A. DeSart. 1999. “Using State Polls to Forecast Presidential Election
Outcomes in the American States.” International Journal of Forecasting 15, 2, 137-142.
Holbrook-Provow, Thomas M. 1987. “National Factors in Gubernatorial Elections.” American
Politics Research 15, 4, 471-483.
Howell, Susan E. and James M. Vanderleeuw. 1990. “Economic Effects on State Governors.”
American Politics Research 18, 2, 158-168.
Jones, Randall J. Jr., J. Scott Armstrong, and Alfred G. Cuzán. 2007. “Forecasting Elections Using
Expert Surveys: An Application Using U.S. Presidential Elections.” Working Paper.
Kaplan, Edward H and Arnold Barnett. 2003. “A New Approach to Estimating the Probability of
Winning the Presidency.” Operations Research 51, 1, 32-40.
Kastellec, Jonathan P., Andrew Gelman, and Jamie P. Chandler. 2008. “The Playing Field Shifts:
Predicting the Seats-Votes Curve in the 2008 U.S. House Elections.” P.S. Political Science and
Politics 41, 4, 729-732.
Kiewiet, D. Roderick. 1979. “Approval Voting: The Case of the 1968 Election.” Polity 12, 1, 170-
Klarner, Carl. 2008. “Forecasting the 2008 U.S. House, Senate, and Presidential Elections at the
District and State Level.” P.S. Political Science and Politics 41, 4, 723-728.
Lacy, Dean and Barry C. Burden. 1999. “The Vote-Stealing and Turnout Effects of Ross Perot in
the 1992 U.S. Presidential Election.” American Journal of Political Science 43, 1, 233-235.
Lewis-Beck, Michael S. and Tom W. Rice. 1983. “Localism in Presidential Elections: The Home
State Advantage.” American Journal of Political Science 27, 3, 548-556.
Lewis-Beck, Michael S. and Tom W. Rice. 1984. “Forecasting U.S. House Elections.” Legislative
Studies Quarterly 9, 3, 475-486.
Lewis-Beck, Michael S. and Tom W. Rice. 1985. “Are Senate Election Outcomes Predictable?”
P.S. Political Science and Politics 18, 4, 745-754.
Lewis-Beck, Michael S. and Charles Tien. 1999. “Voters as Forecasters: A Micromodel of Election
Prediction.” International Journal of Forecasting 15, 2, 175-184.
Lewis-Beck, Michael S. and Charles Tien. 2008. “The Job of a President and the Jobs Model
Forecast: Obama for ’08?” P.S. Political Science and Politics 41, 4, 687-690.
Lewis-Beck, Michael S. and Charles Tien. 2010. “The Referendum Model: A 2010 Congressional
Forecast.” P.S. Political Science and Politics 43, 4, 637-642.
Leyden, Kevin M. and Stephen M. Borrelli. 1995. “The Effect of State Economic Conditions on
Gubernatorial Elections: Does Unified Government Make a Difference?” Political Research
Quarterly 48, 2, 275-290.
Lichtman, Alan J. 2008. “The Keys to the White House: An Index Forecast for 2008.” International
Journal of Forecasting 24, 2, 301-309.
Lockerbie, Brad. 2008. “Election Forecasting: The Future of the Presidency and the House.” P.S.
Political Science and Politics 41, 4, 713-716.
Marra, Robin F. and Charles M. Ostrom Jr. 1989. “Explaining Seat Change in the U.S. House of
Representatives.” American Journal of Political Science 33, 3, 541-569.
Mayer, Jeremy D. 2002. “Nixon Rides the Backlash to Victory.” The Historian 64, 2, 351-366.
Niemi, Richard G., Harold W. Stanley, and Ronald J. Vogel. 1995. “State Economies and State
Taxes: Do Voters Hold Governors Accountable?” American Journal of Political Science 39, 4, 936-
Norpoth, Helmut. 2008. “On the Razor’s Edge: The Forecast of the Primary Model.” P.S. Political
Science and Politics 41, 4, 683-686.
Partin, Ronald W. 1995. “Economic Conditions and Gubernatorial Elections: Is the State Executive
Held Accountable?” American Politics Research 23, 1, 81-95.
Peltzman, Sam. 1987. “Economic Conditions and Gubernatorial Elections.” American Economic
Review Papers and Proceedings 77, 2, 293-297.
Pickup, Mark and Richard Johnston. 2008. “Campaign Trial Heats as Election Forecasts:
Measurement Error and Bias in 2004 Presidential Campaign Polls.” International Journal of
Forecasting 24, 2, 272-284.
Piereson, James E. 1977. “Sources of Candidate Success in Gubernatorial Elections, 1910-1970.”
Journal of Politics 39, 4, 939-958.
Rosenstone, Steven J. 1983. Forecasting Presidential Elections. Yale University Press: New
Rothschild, David. 2009. “Forecasting Elections: Comparing Prediction Markets, Polls, and their
Biases.” Public Opinion Quarterly 73, 5, 895-916.
Sidman, Andrew H., Maxwell Mak, and Matthew J. Lebo. 2008. “Forecasting Non-Incumbent
Presidential Elections: Lessons Learned from the 2000 Election.” International Journal of
Forecasting 24, 2, 237-258.
Simon, Dennis M. 1989. “Presidents, Governors, and Electoral Accountability.” Journal of Politics
51, 2, 286-304.
Soumbatiants, Souren, Henry W. Chappell Jr., and Eric Johnson. 2006. “Using State Polls to
Forecast U.S. Presidential Election Outcomes.” Public Choice 127, 1, 207-223.
Stanley, Harold W. 1988. “Southern Partisan Changes: Dealignment, Realignment, or Both?”
Journal of Politics 50, 1, 64-88.
Strumpf, Koleman, S. and John R. Phillipe Jr. 1999. “Estimating Presidential Elections: The
Importance of State Fixed Effects and the Role of National Versus Local Information.” Economics
and Politics 11, 1, 33-50.
Squire, Peverill. 1989. “Challengers in U.S. Senate Elections.” Legislative Studies Quarterly 14, 4,
Squire, Peverill. 1992a. “Challenger Profile and Gubernatorial Elections.” Western Political
Quarterly 45, 1, 125-142.
Squire, Peverill. 1992b. “Challenger Quality and Voting Behavior in U.S. Senate Elections.”
Legislative Studies Quarterly 17, 2, 247-263.
Svoboda, Craig J. 1995. “Retrospective Voting in Gubernatorial Elections: 1982 and 1986.”
Political Research Quarterly 48, 1, 135-150.
Tompkins, Mark E. 1984. “The Electoral Fortunes of Gubernatorial Incumbents: 1947-1981.”
Journal of Politics 46, 2, 520-543.
Tompkins, Mark E. 1988. “Have Gubernatorial Elections Become More Distinctive Contests?”
Journal of Politics 50, 1, 192-205.
Fundamental Models for Forecasting Elections*
Patrick Hummel David Rothschild
In case of India Mint a business daily published in collaboration with Wall Street Journal promised data driven election analysis article till Lok Sabha elections. One of them is