Wednesday, October 22, 2014

Google and Government: Using Social Network Analysis to Evaluate the Revolving Door Between Washington and Silicon Valley

Miranda Bogen
I will be taking the second module

Project proposal:
Use social network analysis to asses the “revolving door” phenomenon between the federal government and major technology companies, in particular, looking at direction of movement (public-private vs private-public) and the role of professional overlap in determining this movement. Do these moves correlate with social relationships (presumed based on overlap at a federal agency or company, or shared attributes like alma mater) or are these individuals moving independently? 


Hypothesis:
Overlaying key federal staff and their professional history will show that a significant portion of sector-switchers will have overlapped either in government or at a technology company. This would indicate that professional connections are driving this revolving door movement and that networks formed in one sector are either highly or moderately influential in the other. 


Overview:
“Obama and Google Swap Staff”
“Obama & Google: A Love Story” 
“Obama and Google Connected at the Hip”

The concept of the revolving door from the federal government to the private sector and back has traditionally been applied to the telecommunications/media, military contracting, energy, and banking industries. However, in the past few years, major media outlets have reported concerns about the growing “diaspora” of major technology company alumni transitioning into the federal government. Google veterans now occupy top policy advisor positions in the White House Office of Science and Technology Policy and the Federal Communications Commission, and Google and other major technology companies have snapped up White House, Congressional, and campaign staffers to fill communications and public policy/lobbying positions. However, no one has performed a methodical analysis of this phenomenon to determine whether the media hype can be substantiated and what this means for companies, policymakers and regulators.


Selected Timeline of Key Appointments

February 2009: Katie Stanton, Principal of Google’s New Business Development team, joins White House team as Director of Citizen Participation

April 2009: Sonal Shah, former head of global development at Google.org, appointed head of White House Office of Social Innovation

May 2009: Andrew McLaughlin, Google’s head of global public policy, leaves to become U.S. Deputy Chief Technology Officer

January 2013: President Obama appoints Google staffer Vint Cerf to the National Science Board

June 2013: Twitter executive and former Google VP Nicole Wong joins the White House as Deputy Chief Technology Officer; Bloomberg reports that Google hires Obama’s data-mining team

August 2014: Mickey Dickerson, former Site Reliability Manager at Google, appointed to run new U.S. Digital Service

September 2014: White House names Megan Smith of Google X as U.S. Chief Technology Officer; former Twitter council and head of public policy Alexander Macgillivray Deputy CTO

(Google was the third largest corporate supporter of the 2008 Obama campaign, and Google Chairman and CEO Eric Schmidt is also known to have strong ties with the Obama administration and has served in advisory positions.)


Project Plan
To create a data set for this project, I will focus on those individuals who worked at Google and also in the federal government. Individuals will have had to be employed by a federal office during the Obama Administration (since January 2009) and have worked at Google at any point in their career. I plan to infer network connections when an individual shares a workplace with another individual in the same year. To source the data, I will use publicly available information from the Center for Responsive Politics, a revolving door watchdog that documents employment history of registered lobbyists. I will supplement this data with media reports of political appointments and defections, and flesh out professional history and relevant attributes using public information on LinkedIn.

I will create a valued data set to show how many workplace-years each individual shared with another. If this value is more than one, this will indicate that they worked together more than once at either the same or different companies. Based on common knowledge about workplaces and hiring, it can be inferred that these individuals may have used their network to access jobs in a new sector. I will looks a cliques, subgroups, and centrality measures to determine if certain individuals are more influential within the larger group of revolving door employees, and cross-reference with attributes of these notes to determine if such influence can be predicted in the future.

I also plan on looking at the direction of movement between sectors, and delayed but parallel movement by individuals following colleagues in a given direction. If the analysis finds this happening repeatedly, it would indicate that individuals are recommending replacements for themselves as they prepare to move on to different positions, thus shedding additional light on the influence of professional network on career movement between Google and specific government offices.


Conclusion and Possible Future Scope Expansion
Given that both Silicon Valley and Washington are comprised of small and tightly-knit networks, the exploration of network measures in the revolving door phenomenon are likely to reveal dense networks and clear patterns of movements between sectors. This analysis will likely serve to support media claims that the Obama administration has tapped into a pool of technology talent and that these individuals are tapping their own networks to pull former colleagues across sectors. 

This analysis could potentially be expanded to include employees of the top 50-100 technology companies to determine if any trend identified is unique to Google or if the revolving door trend is similar across the technology sector. The outcome of both the preliminary Google analysis and any future extended analysis will be instructive to technology companies looking to understand their ability to influence relevant policy matters, governments looking to recruit technology talent, and to regulators and lobbying watchdogs. 

1 comment:

Christopher Tunnard said...

Very interesting and nicely thought through SNA. My only question is: what's your Question? Shouldn't it be about the qualitative difference (if you find any) between those whose moves are correlated and those that are independent? I know that SNA can't give you the "Answer," but it would nice to make some inferences based on your results. Look forward to watching this progress.