Commentary: Analysis of Senate Votes on the FISA Extension Bill

December 28, 2012. The Senate passed HR 5949 [LOC | WW], the "FISA Amendments Act Reauthorization Act of 2012", by a vote of 73-23. See, Roll Call No 236. The Senate first rejected four amendments on roll call votes.

Voting on amendments, and for final passage, correlated with party affiliation. Senate Democrats were more likely to vote for the amendments, and against the bill, than Republicans.

This is observable by reviewing the roll call votes. In addition, one can quantify this. For example, the Pearson's Correlation Coefficient for the two dichotomous variables of party affiliation (being a Democrat) and voting for the Leahy amendment was .70. Similarly, the coefficients for party affiliation and voting for the Merkley, Paul and Wyden amendments were .64, .18 and .63, respectively. The coefficient for party and voting against final passage was .39.

Party affiliation does not, however, explain all votes. Moreover, turning to data on party affiliation of registered voters in each state does not help. Nor do the ideological ratings of Senators.

One can also sometimes explain voting on the basis of how a particular bill disparately burdens or benefits different states. However, that method of analysis is not useful for this bill because this bill affects all states equally.

This article offers a hypothesis regarding another possible causal factor in Senators' voting decisions on bills of this nature. Senators are also influenced by businesses based in their states, and their contractors, affiliates, and trade associations. The primary businesses involved with electronic surveillance are the communications service providers.

Senators represent geographic territories. To certain extent, communications service providers are geographic too. Consider the 1982 Modification of Final Judgment (MFJ) in US v. AT&T that created seven separate geographically defined Regional Bell Operating Companies (RBOCs). The MFJ did not includes a few states in an RBOC. Over the years six of the seven have been consolidated into Verizon and AT&T.

This article suggests that there was a pattern in voting on the FISA extension bill based upon whether the Senator represents a state that had an RBOC that became part of either Verizon or AT&T, or whether the Senator represents one of the USWest states, or Alaska, Hawaii or Nevada. USWest, of course, became Qwest, which is now part of CenturyLink. For the sake of brevity, this article refers to a dichotomous breakdown of states - USWest Plus states and Verizon/AT&T states.

On this FISA extension bill, voting correlated with affiliations in this dichotomy. That is, the Senators who were more likely to vote for the amendments, and against the bill, represent the USWest Plus states. Conversely, the Senators who were more likely to vote against the amendments, and for the bill, represent the Verizon/AT&T states.

CenturyLink's roots are as a domestic rural land line phone service provider. It has no wireless arm. Also, USWest/Qwest's experiences with federal regulatory, law enforcement and intelligence agencies were dissimilar to those of the other six RBOCs and their successors.

The Pearsons Correlation Coefficient for the two variables of RBOC affiliation (representing a USWest Plus state) and voting for the Leahy amendment was .3. For RBOC affiliation and voting for the Merkley amendment, the correlation coefficient was .31. For RBOC affiliation and voting for the Paul amendment, it was .33. For RBOC affiliation and voting for the Wyden amendment, the coefficient was .34. Finally, for RBOC affiliation and voting against final passage, the correlation coefficient was .40. Notably, being a USWest Plus state Senator was a better predictor of voting than party affiliation on both the Paul amendment and on final passage.

TLJ also scored each Senator on a scale of 0 to 5. TLJ assigned one point for each vote for one of the four amendments, and one point for voting against final passage. The Pearson's Correlation Coefficient for the two variables of party affiliation (being a Democrat) and this composite score was .60. The coefficient for RBOC affiliation (representing a USWest Plus state) and this composite score was .41. Thus, overall, party was a better predictor of voting.

While this article presents some showing of an observable statistical relationship between voting on this bill and representing a USWest Plus state, this does not show causation. It is possible that there is some other factor, that is not discussed in this article, and that is correlated with being a USWest Plus state, that does have a causal affect upon Senators' votes. The correlation coefficients measured for this article results from this other factor. For example, the USWest Plus states are all in the western part of the country. Perhaps there exist regional variations in voter attitudes about government surveillance, with westerners more inclined towards privacy than people in other parts of the country.

(TLJ categorized all Senators as either a Democrat of Republican. Sen. Sanders and Sen. Lieberman were treated as Democrats.)

(Published in TLJ Daily E-Mail Alert No. 2,499, December 30, 2012.)