I usually limit my writing to international relations rather than discuss domestic politics. However, I found the election results quite interesting when viewed through the lens of intelligence studies and the international system. The fact that the Romney campaign appeared to be legitimately surprised by the outcome reveals important patterns about the use and non-use of intelligence for those seeking to understand international relations and policy more generally. In what follows, I operate under the assumption that the Romney Campaign was indeed confident they would win the election, and through a combination of psychological and organizational mechanisms left themselves vulnerable to surprise on election night. It is entirely possible that in time, new materials will come out that show the campaign to have been more reserved and cautious internally than currently believed, and I will also acknowledge that without blinding self-confidence, few candidates would even consider, must less endure a presidential campaign in the first place.
In the immediate run up to the election and its aftermath, the leading story was about Nate Silver, prediction markets, unskewed polls, and pundits that went with their guts only to conclude that no one knew anything.
Interestingly, the work of Silver, prediction markets, and others looking to keep score on the accuracy of different predictions and the analysts and pundits that made them may be the least relevant to serious thinking about policy and politics. My problem with Silver and prediction in general has nothing to do with accuracy but purpose. Simply put, policy makers and campaigns are looking to change the world so forecasts and predictions disconnected from causal narratives run counter to their needs–Silver, prediction markets, and other pollsters all seek to provide a snapshot of the electorate in order to characterize its current state (polling) and likely future (prediction, forecast), while the campaigns are seeking to change the state of the electorate through their actions in order to influence voter behavior and the electorate in general. Thus, assessments that state Obama has a 20% or 80% change of winning reelection are of very limited relevance to the informational needs of the campaigns—which are searching for information regarding how to mobilize and influence voters, frame issues, etc. When election predictions are seen through the lens of intelligence, the difficulties they pose become clear. As Doug MacEachin noted:
Experience has forcefully demonstrated that however much policymakers may differ on political philosophy, they react much the same to intelligence reports that run contrary to their expectations or hopes. Unhappy responses from consumers are a natural part of the intelligence territory. No policy official likes to see intelligence that suggests things are going badly for the policy, any more than a coach likes to hear a prediction that he will lose the game. But the job of a football scout is not to turn in a prediction of the score, but to give the coach information useful in formulating a game plan. The same applies to intelligence analysis. (Douglas J. MacEachin, “The Tradecraft of Analysis,” in Roy Godson, Ernest R. May, and Gary Schmitt, eds., Intelligence at the Crossroads, Brassey’s, 1995, p. 75)
If intelligence analysts are like team scouts, then election forecasters are like Las Vegas bookies seeking to handicap the Super Bowl. In the former, predictions of the game’s final score are not relevant, while in the latter it is paramount.
The irrelevance of predictions, then were matched by the behavior of campaigns, pollsters, consultants, and the media engaged in a self-reinforcing loop of self-deception indicative of so many intelligence failures (although it may be more accurate for failures of this kind to be called policy failures). Here, at least five factors were evident that combined to create a mindset that encouraged personal, organizational, and institutional failure.
First, intelligence professionals have observed from the very creation of the community that policy makers have other sources of information and are free to accept or reject the assessments of analysts as they choose. Indeed, two of the moist poignant observers of the relationship between intelligence and policy, Sherman Kent and Roger Hilsman, each noted that policy makers largely rejected and even became hostile to assessments that challenged their own expectations and positions. Thus, it should be no surprise that the Romney campaign and its backers likely turned towards their own sources while rejecting the independent assessments that proved to be more accurate. In this context, the attacks on Silver are an important part of this defensive pathology and quite familiar to intelligence analysts who often meet skepticism and criticism from consumers when their assessments do not meet policy-maker’s expectations. Because policy makers have multiple sources of information available to them, it should be so surprise that they choose to listen to those that are the most psychologically comforting or affirming of their world views. For optimists, that means sources that tell them they are succeeding, while pessimists may be attracted to doomsday forecasts.
Second, there is a distinct difference between presidential candidates that are looking to influence and mobilize voters and political pollsters looking to characterize the current state of the population. Campaigns exist because of the belief that actions and choices can change the world, and it should be no surprise that the Romney campaign believed they had the power to affect the public, that they were running against a weak opponent, and were confident in their capabilities. This likely reinforced two preferences for information: 1) an innate preference for sources telling them that their efforts were working and that they had closed gaps with voters or even taken the lead, and 2) a focus on identifying causal levers and sources of influence rather than a score card, thus preferring narratives of influence, enthusiasm and action over static public opinion data. Thus, the campaigns and independent pollsters replicated the common cultural pattern of policy makers as optimists that are confident their actions can change the world and independent intelligence producers as skeptics that are for more cautious about the extent to which the will of policy makers and their clever strategies can change things.
A third factor that likely affected the Romney campaign’s resistance to public polling data was the longstanding preference of executives or decision makers to overvalue information from their own sources and discount those from outside of their control. An example of this was demonstrated by General Montgomery’s dismissal of intelligence about the presence of German Panzer Divisions in Arnhem before launching Operation Market Garden, in favor of trusting his own reconnaissance that did not find any evidence of operational German forces. In the case of political campaigns, the repeated references to the campaign’s internal polling were cited as sources of confidence to the candidate, the campaign, and their supporters. It is likely that the existence of internal polls that were reporting favorable results to the campaign reinforced the psychological factors mentioned above, exacerbating the tendency for intelligence consumers to engage in self-deception.
A fourth factor regarding the self-deceptive nature of political campaigns is often seen in intelligence analysis regarding the limitations of independent analysis. Intelligence analysts face pressure to bring policy makers good news and confirmation that their policies will be successful and their commitment to the team is often called into question when their assessments do not comport to expectations. I suspect that this is also a significant problem for internal pollsters, who are likely partisans themselves, or contractors that feel vulnerable if they report to a campaign manager or candidate that they are losing or their strategies are ineffective.
A fifth factor has to do with secrecy and privilege. Many prominent conservatives in the media challenged the public polls and noted that sources in the Romney campaign were confident. What is particularly interesting is that while the campaign was likely deceiving itself, outsiders viewed contact with insiders as a privilege and thus accepted the same deception as a signal that they too had access to secrets known only to a knowledgeable few. Thus, access to the internal polling and narratives that were part of the campaign’s self-deception were ironically prizes to be given to friendly media sources, effectively extending the deception’s reach. This too replicated a common problem where access to secrets is often confused with the accuracy of the secrets themselves.