Actionable Intelligence and Prince William County
In the general elections of 2008 Virginia voted democratic for the first time since 1964 with Obama carrying the state. Demographic shifts and increased voter participation rather than a shift in political allegiances account for this outcome. This demographic shift was an artifact of the dynamics of capital accumulation, uneven development and class struggle in a period of neoliberalization, prefiguring zoning and land use patterns.
Focusing on Prince William County, Virginia, I applied spatial interpolation techniques in a GIS to translate the 2008 election returns from the geography of precincts to year 2000 zoning classification areas for further quantitative analysis. The goal was to produce actionable intelligence for working class organizations building popular power at the base. The results suggest that the economic logic of the micro-politics of land use and zoning prefigure political change at larger scales of aggregation.
This project consisted of two parts: an analysis of the 2008 election returns for Prince William County Virginia and a spatial representation of voting records for the 111th congress using Geospatial methods to analyze and present the returns cartographically. The logic of land use and zoning are independent from the logic of the hierarchy of census enumeration units and political precincts, masking the impact of class and power relationships at the level of a single county.
Land use decisions determine the density and type of housing and pattern of economic development in a given area, in turn shaping the characteristics of the population and electorate. Actions at this lowest level jump upward to the national scale in the process of redistricting where the boundaries for congressional districts are defined at the state level. I use several geospatial methods: areal interpolation to transfer values from one set of units to another and point pattern analysis to measure clustering of congress people. By making the geography of election outcomes transparent as attributes of land use and housing density polygons, housing density itself becomes an independent variable to predict voting behaviors.
At the national scale, congressional voting records are converted to a set of x, y coordinates and scrutinized using hierarchical clustering, classified using natural breaks and grouped as binary categories to generate strategic insights. A selection of only Virginia Congress members are visualized, revealing the strategic importance of redistricting and civic engagement projects in light of demographic shifts linked to zoning and land use decisions on the micro-scale of local politics. The results are presented as maps and diagrams which might illuminate challenges and opportunities for organizations engaging with electoral efforts.
Part I Prince William County
Many find the brave new globalized world to be threatening and alien to their imaginary of place and nation, but are unwilling to question its underlying assumptions or causes, such as neoliberal global restructuring. One could speak of a politics of fear directed toward a dehumanized Other. In Prince William County (PWC; see Figure 1), this fear drove implementation of draconian anti-immigrant legislation in 2006, prefiguring similar initiatives proposed at municipal, county and state levels nationwide such as the recent legislation enacted by the State of Arizona. The specific dynamics seen in Prince William County might be playing out across the nation as the Westphalian certainties of the nation state as a discrete entity are superseded by a more complex set of spatial relationships. New political, social and economic spaces such as global trade networks and free trade zones are challenging the nation state as exclusive site of sovereignty and citizenship at the global scale, while at the same time privileging cities and urban regions in an emerging neoliberal order (Harvey 2005). A nativist reaction to these trends therefore confronts a social landscape shaped by local and global trends – zoning and land use patterns, global economic restructuring and neoliberalization—which might not be reversible by merely sealing borders or enacting punitive legislation.
The economic logic of the micro-politics of land use and zoning prefigure political change at higher levels of aggregation. A quantitative analysis of the 2008 election results and zoning patterns in PWC demonstrates what activists intuitively grasp: that working class people of color are found in high density housing. But, that high density housing is also home to a multi-national mix.
Figure 1 Prince William County Virginia and Washington DC
A key dimension of the shift from the embedded liberalism of fordist production to the neoliberal regime of free market liberalism and global production chains was privileging of the new economy of high technology, specialized services and finance (Dicken 2003, Hackworth 2007, Harvey 2005, Smith 1982, Sassen 2000), itself dependent on massive state intervention based on defense spending and Pentagon contacts (Markusen et al. 1991; Beauregard, 2006). Northern Virginia is a prototypical example (Dennis 2009).
Since the onset of neoliberal restructuring, tendencies towards dispersal have accelerated across the Washington Metropolitan area. Employment centers like those in Prince William County (PWC) are led by the high technology and producer services and are clustered. The regional is networked, small clusters of service producers are linked together with the central business district in Washington being but one area of concentration.
Changes over the last half century from exclusionary residential zoning, de jure segregation, and since 1965, relatively relaxed immigration laws have contributed to the emergence of hetero-local settlement patterns and ethnoburbs across the nation. As well as being magnets for highly educated workforces of the information society, developing urbanized cores like those found in such places as PWC in Northern Virginia are also attracting low skill workers in a global labor market. The Washington area and in particular Northern Virginia is a “New Immigrant Gateway”( Price, Cheung, Friedman Singer 2005) giving rise to contentious politics and conflicting images of what the region is and what it will become.
GIS has been the subject of a heated debate concerning the embedded ontologies, epistemologies and politics of techno-science (Pickles 1997). Critics see the technology as inherently an exemplar of instrumental rather than critical reason. I however assert that the techniques and tools of quantitative analysis and spatial analysis can be framed and deployed from a critical perspective. Here, GIS is a tool used to expose the hidden geography of power through the use of interpolation to extract values from one geography and transferring those values to another. The geographies of election precincts and zoning are independent of each other and are not directly comparable. A solution to this problem can be found in interpolation, converting election returns into a surface (figure 2) and extracting values onto the zoning features (fig4) to produce a table for further analysis. Zoning classifications are at the nominal level of measurement[1] while election outcomes can be made dichotomous, reducing the data to a 2 by 2 table.
Interpolation relies on the theory that things closer together are more similar than things farther away a regularity often referred to as Tobler`s “first law of geography” (Tobler 2004), a regularity without which no geography would be possible. The centroids for zoning area polygons were used as sampling points for the values of the interpolated surface (figure 3) and added onto the attribute table for the zoning feature layer for further analysis. This interpolation technique could be used to extract values from other incompatible areal units for multivariate analysis.
Privatization has led to the displacement of federal employment and stable government workforces by precarious contract work at both the highly skilled and unskilled levels(Dennis 2009), particularly with regards to military and defense related industries, comprising a social and political mass beyond narrow definitions of class race and gender, as inhabitants with a “right to the city”, “The right to the city stresses the need to restructure the power relations that underlie the production of urban space, fundamentally shifting control away from capital and the state and toward urban inhabitants”. (Purcell 2002:101) That mass is made up of precarious workers, both high-pay high tech and low pay-low tech service workers that come together as urban inhabitants with distinctively urban interests as suggested by the 2 by 2 contingency table for high and low density zoning areas and Obama support in Prince William County (Table1). There were only two variables: housing density and Obama margins (I subtracted percentages for McCain from percentages for Obama and converted the difference to absolute values by winner). Table 1, shows χ2 (1df, N=873) = 100.45, p<.001 meaning that the 2008 election results were strongly correlated to zoning density. High density residential concentration alone without considering income, race or gender was related to high Obama turnouts. These areas might contain expensive or affordable housing the common denominator for all of these areas is urbanization. In this election, it was this urban subject that was most responsible for the Democratic victory.
Neoliberal global restructuring in addition is creating new forms of belongings and “reconfiguring relationships between governing and the governed, power and knowledge, and sovereignty and territoriality” (Ong 2006:3).The Nativist and the Tea Party movements might be trying to rebuild the Reagan coalition around right wing neo-populist ideology: taxpayer/homeowner/ethnic nationalism (read white) to counter a left-leaning multi-national urban mass that put Obama in the White House. This urban mass, an emerging multi-national multicultural new majority is constitutive of neoliberal urban restructuring when considered in relation to other changes. Regardless of the current political climate or individual election outcomes these changes indeed might be resistant to efforts designed to retrench or roll back reforms.
Figure 2 2008 election results by precinct in Prince William County, VA
Figure 3 Conversion to an Interpolated Surface (see text for discussion)
Figure 4 Areal Interpolation Showing the Margin of Victory for McCain (red) and Obama (blue) in the 2008 Presidential Election (see text for discussion)
Table 1generalized contingency table for zoning categories and interpolated election values
|
winner |
Low density |
High density |
total |
|
Obama |
17.87% |
49.37% |
581 |
|
McCain |
20.27 |
13.17 |
292 |
|
total |
37.46 |
62.54 |
873 |
χ2 (1df, N=873) = 100.45, p<.001
Chi-square (Pearson's) =100.45 Fisher's P=2.06, p<.001
The 111th Congress
At a level and a venue much removed from Prince William County, voting records of the US House of Representatives reveal patterns similar to those apparent at the more intimate level of one urbanizing county. Figures 5 and 6 show the political composition of the 111th congress as ranked by the "That’s My Congress" website: http://thatsmycongress.com/house/ . On these figures the two axes show regressive (y) and progressive(x) voting records. These scores are independent from each other,values for three indexes of core principles including freedom, knowledge, and an economy of opportunity, peace and environmental security supported or opposed were created based on votes on specific bills and concerted to percentage scores for each congress person. Scores were normalized for each congress person by converting them to Z scores so that the center of the x and y axes would be the exact political center based on all the scores both regressive and progressive. My interpretation of this diagram (figure 5) is counter-intuitive to the perception of a disorganized and contentious democratic caucus and a focused and united GOP. I found diverse right with a focused and tight left of center block dominated by a smaller bloc of center-right democrats trying to align with the more moderate republicans against the left.
Congress was polarized as illustrated by the values in each of the four quadrants (Figure 5). The "left" dominates congress with 202 members, while the "right" has just 179 members. Two maps are shown here, one for the entire congress (Figure 5); one for just the VA delegation (Figure 6). The VA map has a shadow grid showing the relative position of other members of congress. It is a Triangular Irregular Network (TIN) tracing out the positions of the state delegation against congress as a whole. A TIN is a particular triangulation where lines connecting points do not cross and the angles of the triangles formed by those lines are less or equal to 90 degrees(O`Sullivan,Unwin 2002:224).The lengths of lines and connectivity between the locations of points representing congress people could be used as proxy measures to quantify ideological relationships between members and the relative distances between them.
Voting patterns and party do not align exactly; congress is split by party with 255 democrats and 178 republicans with 12
members whose voting records are not predictable by party affiliation; split evenly between Democrats and Republicans as shown by a discriminant analysis. Discriminant analysis is a
statistical method used to “predict group membership from a group of predictors” (Tabachnick, Fidell 2001: 456). Discriminant analysis is similar to ANOVA but in effect turns the variables
around, using the group membership as the dependent variable and the predictors, in this case progressive and regressive voting records as the independent variables. Here the objective is to test
the reliability of combining progressive or regressive voting patterns to predict party affiliation reliably. A discriminant analysis showed however that party affiliation and voting patterns do
align 97% of the time; the exceptions are listed, congress persons who vote independently. These twelve congress people and the seats they occupy might represent what might be called the
bipartisan block in a congress where the moderate right as suggested by the upper right quadrant (figure 5) only contains one member Rep. Mark S. Kirk (R-IL 10). The center or the lower left
quadrant has 51 members, the controlling block of the 111th congress.
Congress persons that appear below the X axis vote consistently against regressive legislation, while those to the right of the y axis vote consistently for progressive bills. The outliers are Rep Maurice Hinchey (D-NY 22) on the left and Rep Trent Hanks (R-AZ 2) on the right.
The left, the core of congress, is clustered with 154 congress persons consistently voting in similar ways, they are in the
lower right quadrant much more progressive than the center (lower left quadrant) where the speaker of the house lies. The house members are color coded by progressive scores. The class
intervals for each of the five color bands are based on natural breaks, meaning that upper and lower values for each band are determined by natural groupings not by arbitrary values. The
most conservative group is red, shading to dark blue for the most progressive. The color bands are independent of party. The right is less unified than the left, even if it contains the
largest number of house members (165) The trend diagram illustrates that the spread of right wing congress persons is wider than the left, something also evident in the larger diagram (Figure 5)
but where it is not so clear because many house members have exactly the same scores on the x and y axes. This suggests that a vigorous progressive partisanship rather than a search for an
elusive center might be an effective political strategy (Krugman) when centrists of the right and left (lower left and upper right quadrants) only account for 52 house members.
The center and center right is the next largest grouping (orange) with 95 members. The last three classes while individually small, never the less contain 173 members, the far left (dark blue) being the smallest sub-group, with only 19 members.
The insert cartogram shows congressional districts exaggerated to reflect the size of their population and colored according to
the progressive score classes. Districts with progressive representatives tend to be located in the urban NE and on the west coast with outliers in the south and mid west. Center and center right
members are more likely to come from suburban and rural areas containing urban clusters. Right wing congressional districts are large and rural in many cases.
The bottom line is that while polarized, right wing house members are more dispersed in their voting behavior than the left. The center is smaller than either the right or left but holds the balance of power even though the core of congress (154 house members closely allied in voting patterns) is left.
Figure 6 shows just the Virginia delegation. The Virginia delegation is polarized with an even sharper division between left and right than the larger body. The insert cartogram shows districts sized according to population, visually it appears that Virginia might be more progressive than its popular image seems to propose.
Figure 5 The 111th Congress
Figure 6 The Virginia delegation only
By contesting neoliberalization at multiple scales, the limitations of any particular scale of engagement could be mitigated. Praxis, like neoliberalism is path dependent. It must be specific to a particular conjuncture so action will cut through scales, facilitated by thick organic trans-local networks operating at multiple scales across a complex geo-political and economic landscape. By creating narratives of resistance, connecting here and there in a project to expand and reclaim the commons, actions taken at the lowest levels at a key node in the circuits of power might resonate at higher levels.
While elections can be won or lost, it is doubtful that the transitions set in motion by neoliberal restructuring can be undone. The United State will become a majority minority nation—a nation where euro-Americans are but one minority among minorities—regardless of strict border control or draconian anti-immigrant measures enacted at the local level. In this context, the legitimacy of the state and state institutions depends on a civic rather than ethnic construction of national identity and a restoration of a welfare state consistent with national democratic pretensions.
In a rush toward post modernism and a rejection of all things positivist, the methods of spatial analysis have been rejected. Spatial analysis however can be detached from its positivist moorings since many geostatistical procedures are inherently subjective and dependent on the ontological and epistemological understandings of the operator. The digital world is a data rich environment, much of which includes spatial references. A GIS allows diverse data sets to be visualized and analyzed making complex and diverse information comprehensible. Geospatial technologies can assist in strategic planning process and can be used to create roadmaps to change without predetermining a destination. As actionable intelligence the view from above complements local.
Knowledge. In a GIS, secondary data is twisted and tweaked to reveal and visualize patterns consistent with critical epistemologies and ontological perspectives. Maps charts and graphs as inter-textual representations can be palimpsests and over written. Maps are interactive, the map user defines the route and the meaning of spatial relationships expressed cartographically, generating new insights and new theory to guide political engagements.
References
Beauregard, R. (2006)”When America Became Suburban”. University of Minnesota Press: Minneapolis
Dennis, M (2009) “The New Economy and the Modern South” Gainesville: University of Florida Press
Dicken, P. “Global Shift: Reshaping the global economic map in the 21st century” (2003) NY. Guilford Press.
Hackworth J. (2007) “The Neoliberal City: Governance, Ideology and Development in American Urbanism”, Ithaca, NY: Cornell University Press
Harvey, D. (2005) A Brief History of Neoliberalism. University of Chicago
Levine, N. (2009) CrimeStat III: A Spatial Statistics Program for the Analysis of Crime Incident Locations (update version 3.2a). Ned Levine & Associates, Houston, TX, and the National Institute of Justice, Washington, DC.
Markusen, A.; Hall, P.; Campbell, S.; and Deitrick, S. 1991. “The rise of the Gunbelt: The military remapping of industrial America”. New York: Oxford University Press.
Ong, A.(2006) “Neoliberalism as exception: mutations in citizenship and sovereignty” Durham, NC: Duke University Press
Price, M; Cheung, I; Friedman; S; Singer A. (2005), “The World Settles In: Washington DC as an Immigrant Gateway”. Urban Geography 26(1) 61–83.
Pickles, J. (1997) “Tool or Science? GIS, Technoscience, and the Theoretical Turn” Annals of the Association of American Geographers. 87 (2) 363-372
Purcell, M. (2002) “Excavating Lefebvre: The right to the city and its urban politics of the inhabitant,” Geojournal, 58(2-3) 99-108.
Sassen, S. (1991) “The global city: New York, London, Tokyo” Princeton, N.J.: Princeton University Press.
Smith, N. (1982) “Gentrification and uneven development”. Economic Geography. 56(2) 139-155.
Tabachnick, B.; Fidell, l. (2001) “Using Multivariate Statistics” Fourth Edition. Boston: Allyn and Bacon
Tobler, W. (2004) “On the First Law of Geography: A Reply” Annals of the Association of American Geographers. 94 (2) 304-310
O`Sullivan, D.; Unwin D.(2002) “Geographic Information Analysis” Hoboken: John Wiley and Sons Inc.