We are pleased to welcome professors Charles Tien of Hunter College and Michael Lewis-Beck of the University of Iowa with what we hope will be a regular feature on The Monkey Cage over the next 12 months: their current “nowcast” of the 2012 presidential election.

In contrast to the usual election forecasting approaches, we offer nowcasting.   An electionnowcast predicts what would happen if the election were held “now” (Lewis-Beck, Nadeau, Bélanger, 2011).  Thus, the nowcast acts as an invaluable early warning device, signaling what will come to pass unless things change.   The nowcast comes from a relevant statistical model, whose parameter estimates are held valid across current moments.    That is, the model (with its fixed constant and slope values) predicts the election outcome based on current (changing) X values, as the final contest approaches.  Nowcasting, then, is dynamic, and election predictions may be issued on a quarterly, monthly, or even daily basis, with updates until the actual election occurs.  For example, a nowcast of the November 2012 US presidential election can be issued “now” (November 2011) twelve months before the actual election , in December 2011 (eleven months before), or in January 2012 (ten months before), and so on to the eve of the actual contest itself.

 

To illustrate, our contemporary nowcasts (of about one year before the election) predict a narrow victory for Obama:  the November 2011 nowcast = 51.0 percent for Obama, the December 2011 nowcast = 51.9 percent for Obama.    We will continue to issue further monthly nowcasts until the election has past, to signal the increasing (or decreasing) likelihood of an Obama victory. Below, we explicate the theory and data behind our nowcast equation, which we label a proxy model.

 

THEORY

US presidential election forecasting models abound (see the recent review in Lewis-Beck and Tien, 2011).   Virtually without exception these models base themselves on substantive explanations of the presidential vote.  Our nowcasting approach, however, does not depend on a substantive model. Instead, it rests on identification of a variable that proxies the presidential vote share.  Through location of such a proxy, precise prediction of that vote share becomes possible.   Proxy variables are standard econometric fare.  If a variable is unobservable, as is a future election outcome, then a proxy for it may be sought.    For the proxy to be a good one, it must correlate highly with the unobserved variable.   In a forecasting context, that means the proxy must approach empirical redundancy with the variable proxied.  One successful example of a proxy model in presidential forecasting comes from the French case (Nadeau, Lewis-Beck, and Bélanger, 2011).  Here we use a proxy model as the basis for our nowcasts of the US presidential election outcome.

 

PRACTICE

Our proxy notion expresses itself in the following equation,

Vote = ƒ (Vote Proxy).                                                   (eq. 1)

where Vote = the two-party popular presidential vote share; Vote Proxy = an observed indicator of the unobserved vote.  We offer as a proxy the National Business Index (NBI), yielding

Vote = f(NBI)                                                                (eq.2)

where NBI = the percentage of respondents who say “business conditions are better” minus the percentage of respondents who say “business conditions are worse,” as measured in the national University of Michigan Survey of Consumers.  This variable shows itself to be quite sensitive to actual US business conditions.  For example, in the Great Recession year of 2008, it achieves its maximum negative value, at -81 (meaning overwhelmingly see worse business conditions).  At the other extreme is its maximum positive value, at +47, registering the prosperous year of 1984.

This NBI, measured in April six months before the November election, correlates highly with incumbent vote share, r = .83.  Figure 1 shows this strong link, in a scatterplot over the election sample, 1980-2008.   From a forecasting perspective, the six-month lead time offers considerable advantage.  Firstly, its high accuracy comes at an impressive distance from the election itself, far from the trivialities day-before (or month-before) forecasts offer.  Moreover, this longer lead performs better empirically than the commonly used three-month lag (r = .70).   The strength of the six month lead comes as no surprise.  Other forecasting work, in the US, UK, and France suggests it is in fact optimal; Lewis-Beck and Rice, 1992; Nadeau, Lewis-Beck, and Bélanger, 2010;  Gibson and Lewis-Beck, 2011).

The regression estimation (ordinary least squares) of this Proxy Model yields the following:

Vote = 51.67* + .09* NBI t-6 + e                              (eq.3)
(48.23)   (3.69)

R-sq. = .69, Adj. R-sq. = .64,   SEE = 2.96   D-W = 2.45,  n = 8,

where the variables are defined as with eq. 2; the asterisk indicates statistical significance at .05, two-tail; the figures in parentheses are t-ratios; the R-squared = the coefficient of multiple determination; the Adj. R-squared = the R-squared adjusted for degrees of freedom;SEE = the standard error of estimate; D-W = the Durbin-Watson test; n = 8 observations on US presidential elections 1980 -2008.

The Proxy Model of eq.3 has promising properties.  In addition to the fit statistics (of the R-squared, the Adjusted R-squared, and the SEE), it is worth examining the error produced from predicting the individual elections under study.  Here are these within sample errors:

1980 = -1.06, 1984 = 3.49, 1988 = 1.33, 1992 = -1.61, 1996 = 2.39; 2000 = -5.08; 2004 = -1.02; 2008 = 1.56.

Note that the most extreme NBI year, 2008 = -81, is still predicted with little error (under 2 points).  Further, we see that, with the exception of 2000, the winning party is correctly predicted, though one could argue that the 2000 popular vote winner was correctly predicted.  Moreover, the exceptional case itself encourages acceptance of the model, since virtually all forecasting efforts for the Gore vs. Bush contest were well off,  (LewisBeck and Tien, 2001).  In addition, a healthy distribution of residuals can be observed, with four positive and four negative signs.  Overall, we see the mean absolute error (MAE) is only about 2.19 points.  And, if the series is trimmed, by removing the curious 2000 case, the MAE falls to 1.78 points.  These results suggest that the model can pick the winner in all but the closest races.

A final diagnostic is that its fit could not be improved, despite the addition of other obvious variables, i.e. incumbency and presidential popularity, at different lags.  In particular, it should be noted that presidential popularity fails to add significantly to the model, due to its high collinearity with NBI e.g., r = .85 in October.  The interesting implication is that the effects of popularity are absorbed by NBI.  As well, the effects of the macro-economy appear transmitted by NBI, e.g., for the correlation of NBI and GNP growth, r = .68 in October.  (Nadeau and LewisBeck, 2001, show that NBI outperforms standard macro-economic measures in predicting presidential vote support).  One implication is that the strong correlation between presidential vote share and NBI is far from spurious.  Instead, NBIsucceeds in empirically capturing fundamentals that operate on voters, such as presidential popularity and the macro-economy.

 

NOWCASTS FROM THE PROXY MODEL

The Proxy Model has several desirable properties as a forecasting equation: complete parsimony, long lead time, ease of replication and good model fit (Lewis-Beck, 2005).  In the nowcasting context, we simply apply the proxy model as if the election were upon us – this month, or next month, or the month after – so generating a series of nowcasts stretching to the election itself.  To begin, let us employ it to generate a current month nowcast, assuming the current month is November 2011.  Then, the equation estimates the (November 2011) election outcome from the NBI six months before (April 2011), as follows:

 

Vote _{nov 2011} = 51.67 + .09 (-8)                      (eq.4)

= 50.95 Nov 2011 nowcast

 

where the prediction equation is as with eq.3, and – 8 is the NBI for April 2011 (indicating that more people thought the economy “worse” than “better.”).  As a further example, consider the nowcast for the subsequent month of December 2011 from the May 2011 NBI (when slightly more respondents thought the economy was “better” than thought it “worse”):

 

Vote _{dec 2011} = 51.67 + .09 (2)

= 51.85 Dec 2011 nowcast.

 

CONCLUSION

On the basis of nowcasts, about one year away from the actual election, President Obama looks like a winner, although the margin of victory will be quite small.  These results provide an early warning signal to Republicans, implying that unless things change they will not occupy the White House this coming fall.  Of course, things can still change, and those changes will be tracked in our subsequent nowcasts as the months pass, and we move closer to the actual November 2012 election.  We plan to offer monthly, up-to-date nowcasts on these pages, until the election arrives.  Of course, as we approach April 2012, which affords the NBIdata four our “final” prediction, confidence in these nowcasts will increase.  Will Obama hold his lead?  As the months unfold, we shall see.

REFERENCES

Rachel Gibson and Michael S. Lewis-Beck, 2011. “Methodologies of Election Forecasting: Calling the UK ‘Hung Parliament,”  Electoral Studies,Vol.30(2),June, pp.247-249.

Michael S. Lewis-Beck, Richard Nadeau and Eric Bélanger. 2011. “Nowcasting v. Polling: The 2010 UK Election Trials,” Electoral Studies,Vol.30(2), pp.284-287.

Michael S. Lewis-Beck, Eric Bélanger and Richard Nadeau. 2010. “Election Forecasting in France: A Multi-Equation Solution,” International Journal of Forecasting, Vol.25:1,pp.11-18.

Michael S. Lewis-Beck.  2005. “Election Forecasting: Principles and Practice,” British Journal of Politics and International Relations, Volume 7,No.2,pp.145-164.

Michael S. Lewis-Beck and Tom W. Rice.  1992. Forecasting Elections, Washington, DC: CQPress.

Michael S. Lewis-Beck and Charles Tien.  2001. “Modeling the Future: Lessons from the Gore Forecast,” PS: Political Science and Politics, Vol. 34, No.1,pp.21-23.

Michael S. Lewis-Beck and Charles Tien. 2011.“Election Forecasting,” in The Oxford Handbook of Political Methodology, eds., Michael Clements and David Hendry, chapter 24, pp.655-671.

Richard Nadeau and Michael S. Lewis-Beck.  2001. “National Economic Voting in U.S. Presidential Elections,” Journal of Politics, Vol. 63 (No.1,February), pp.159-181.

Richard Nadeau, Michael S. Lewis-Beck and Eric Bélanger.  2011.  “Proxy Models for Presidential Election Forecasting; The 2012 French Test,” French Politics, forthcoming.

P. took the subway to Bowling Green. On his way to the exit, he passed a line of police officers accompanied by bomb-sniffing dogs. Outside, police had surrounded the “Charging Bull” with barricades and, a few blocks north, sealed off a stretch of Wall Street around the Stock Exchange. P. tried to look nonchalant as he carried a black messenger bag that contained a first-aid kit, a bottled solution of liquid antacid and water (to remedy the effects of tear gas and pepper spray), fifteen Clif bars (carrot cake), and several hundred photocopied maps, showing seven possible locations. “We decided that low-tech communication methods would be best,” P. told me. “If we’d used a mass text message, or Twitter, it would have been easy for the police to track down who was doing this…”

I think the press is way too focused on media strategies — both as they say in the business paid media and earned media — and way too little on grassroots organizing and the so-called “ground game” of politics. Interest groups get under-covered tremendously. There’s also kind of moralism in political journalism; that there are good guys and bad guys; that people are being tested on character; that they are being caught doing bad things or are innocent of doing bad things. There’s a tendency not to understand larger forces — to use a kind of “great man theory” of history — and not to understand politics in the way that political scientists generally do: as a realm where interests come to contend and try to run societies either peacefully or not. Interest groups tend to be treated as illegitimate actors. Compromise tends to be undervalued. Legislation tends to be undervalued. Within political coverage, there tends to be too much focus on the executive branch and not enough on the legislative branch.

Lavine and his colleagues designed an online survey and got responses from a sample of about 800 citizens, including many who expressed sympathy for the Tea Party and many who did not. The survey asked about programs designed to help people who can’t keep up with their mortgage payments stay in their homes…

See here. (Run it through Google translate if, like me, you don’t know any Russian.) I don’t know anything here, will defer to the experts on this one.

“Record High Anti-Incumbent Sentiment,” Gallup reports.  Here’s a graph:

Finally (at least for today), we present the following response to the Russian parliamentary elections from Andrew Little, a Ph.D. candidate at NYU who is writing a dissertation on noncompetitive elections. In response to my queries, Andrew offered the following six points in response to the 2011 Russian elections:

1. Noncompetitive elections—those where the ultimate winner is not in doubt—matter. In one of my papers, I argue the main reasons they matter is because of the information they generate. Even if United Russia was never in danger of losing control of the Duma, the results last week seem to have drastically changed beliefs about the strength of Putin/United Russia and how long the regime will last. These elections are not window dressing or a facade to confer legitimacy—by what definition of legitimacy could anyone possibly be convinced by these elections that Putin’s rule is legitimate?—but are meaningful political events because they generate public information about the strength and popularity of the incumbent and opposition groups.

Most of the analysis of this election has not been about the implications of United Russia no longer having a supermajority in the Duma, but about the information generated by the lower-than expected result. I think this is the proper thing to focus on in Russian elections and in other noncompetitive elections.

2. Fraud is not necessarily about winning elections, but is often an attempt to manipulate the information generated by elections. While this election got pretty close, United Russia consistently has cheated in elections that were not close. In fact, fraud is rarely pivotal in determining the winner of elections and there is more fraud in elections that are not close (See the work of Alberto Simpser ). Joseph Kennedy may have claimed to be “willing to buy as many votes as necessary to win, but he was damned if he would buy a single extra one,” but this sentiment is inconsistent with the empirical record on fraud.

3. Even if fraud is about distorting information, it may not fool anyone. Much that is written about fraud conjures images of powerful dictators manipulating passive citizens and outside observers into seeming invincible, but these observers are clearly well aware that fraud is going on. A great quote along these lines comes from Sergei Kovalev, a Russian Democracy advocate: “You lie, your listeners knows this and you know that they don’t believe you … Everybody knows everything. The very lie no longer aspires to deceive anyone, from being a means of fooling people it has for some reason turned into an everyday way of life, a customary and obligatory rule for living.’’

4. So why does fraud happen? In a working paper I argue that since fraud is a (partially) hidden action, incumbent leaders can’t “commit” to hold completely honest elections. Observers know this and as a result will always infer there is going to be fraud, so incumbents not committing fraud would seem weaker than they really are. In game-theoretic terms, fraud doesn’t fool anyone in equilibrium, but since committing more or less than expected can fool people it still occurs.

5. It may seem odd given the cheating we observed that United Russia allowed international monitors like those from the OSCE, surely knowing that their report would at least be somewhat negative. However, imagine what would have happened if United Russia banned all international monitors. I suspect that citizens, opposition groups, and the international community would have just assumed that there was even more cheating, and the end result of the election demonstrating the unpopularity of the regime would have been unchanged. That is, the official result may have changed, but not the information conveyed by the result.

6. Protests over fraud may be less about fraud than the fact that the election result revealed the weakness of the regime. Most of the protests seem centered around the regime cheating, but the regime has been cheating for a long time. What has changed is that they did less well in the election, signaling weakness and potentially lowering the costs or increasing the benefits to protest. So the fact that protesters think fraud was committed matters, but only in the sense that it means for a fixed reported election result, increasing beliefs about how much fraud was committed makes observers thing the regime was weaker or less popular.