The Supreme Court's denial of certiorari in the same-sex marriage cases can be explained in many different ways. I explain it as a makeshift or workaround. It is a substitute for a familiar method of constitutional change-- getting a supermajority of individual states to support change by state judicial or statutory reforms, followed by the Supreme Court's ratification of the change at the federal constitutional level. This wasn't possible in the case of same sex marriage. So instead the Supreme Court has used the lower federal courts to create a simulation of a bandwagon effect in the states, creating facts on the ground to which it can then later defer.
A familiar method of constitutional change-- call it the "supermajority-of-states" method-- uses the American federal system as a benchmark for legitimate change. It asks if there is a consensus of states in favor of constitutional change. If there is a bandwagon effect in favor of change at the state level, federal courts-- and the Supreme Court in particular--feel more comfortable in confirming the trend in federal constitutional doctrine. This phenomenon is sometimes described as the Court "policing outlier jurisdictions," but as Justin Driver has recently pointed out, the process can occur in many different ways and for any number of reasons.
The supermajorty-of-states strategy leverages the fact of fifty different state jurisdictions with their own laws and state constitution to allow gradual accretion of support for constitutional change at the federal level. Lower federal courts will obviously get in the game earlier than the Supreme Court, but the point is that the Supreme Court will generally wait to confirm a trend in the states until (at least in the Court's eyes) it is both clear and overwhelming.
This strategy was not available in the case of same-sex marriage because of the intervention of state mini-DOMA laws between 2004 and 2008 that, in most cases, entrenched prohibition of same-sex marriage in state constitutions. To be sure, state legislative and judicial change was possible in the remaining jurisdictions, but taken together these remaining states would not constitute anything close to a majority, much less a supermajority.
This was an extremely clever move on the part of marriage equality opponents-- they understood that engaging in preemptive strikes against same-sex marriage in a majority of states would mean that marriage equality advocates would have to go to a federal forum *without* being able to establish a bandwagon effect in the individual states. The strategy sought, in other words, to prevent marriage equality advocates from using anything like the supermajority-of-states approach to constitutional change.
There are, however, alternative methods of legitimating change, namely national popular opinion and elite opinion. Both of these swung decisively toward same-sex marriage in a relatively short period of time. Even many elites in the Republican Party quietly supported for same-sex marriage, and as national public opinion changed, increasing numbers of Republican political elites began to express their views in public. Once political elites in both parties began to support same-sex marriage, a new pathway emerged.
The Supreme Court (or rather, the five Justices who supported same-sex marriage) engaged in a two-step process. First, the Court struck down the federal Defense of Marriage Act in an ambiguous opinion in Windsor, while refusing to decide the larger question of the constitutionality of state bans on same-sex marriage in Hollingsworth. Then the Court let lower courts decide what Windsor meant in the context of changing popular and elite opinion that was increasingly in favor of same-sex marriage.
The results in the lower courts were overwhelming. Virtually all lower federal courts that considered the question held that bans on same-sex marriage were unconstitutional. Most of these courts, although citing Windsor, did not try to precisely match its ambiguous reasoning.
The came the second step: The Court denied cert in the lower federal court decisions. The result is that same-sex marriage is now protected-- or will soon be protected--in a supermajority of states-- around 30 or so.
This two-step process-- an ambiguous decision in Windsor followed by a series of cert denials--compensates for the effects of political mobilization by same-sex marriage opponents in the early 2000s. The cert denials create a supermajority of states protecting same-sex marriage that could not have been achieved by working state by state through state courts and legislatures. It creates facts on the ground that cannot easily be undone. The effect is achieved, however, not by a series of *state* decisions to legalize same-sex marriage but as a result of lower federal court decisionmaking signaled by Windsor. The two-step process does not leverage a bandwagon effect among the *states*; rather it uses a bandwagon effect *among lower federal courts* to create a supermajority among states, which then ensures the ultimate outcome in the Supreme Court.
The result of the cert denials is that federal constitutional protection of same-sex marriage nationwide is almost inevitable. With so many states now guaranteeing same-sex couples the right to marry, it seems impossible for the Supreme Court to turn the clock back and deny same-sex marriage in most states. And if a court of appeals ever rules against marriage equality, creating a split in the federal circuits, it will be easy enough for the Supreme Court to bring in the remaining jurisdictions.
Nevertheless, the remaining jurisdictions that have not adopted same-sex marriage are not "outliers" in the standard sense of that term. They are not outliers because all of the other states willingly sought marriage equality (although many did!). The remaining jurisdictions are outliers because they didn't happen to be in federal circuits where lower federal courts found for marriage equality.
There is considerable debate in the law reviews about whether Windsor is or is not a federalism decision. At this point it seems hardly relevant. Windsor has proved to be a way to get *around* federalism-- that is, around preemptive politics at the state level by marriage equality opponents. The cert denials allow the Court (or a majority of the Court) to say: "Look at the facts on the ground. Look at how many states now protect same-sex marriage. We can't turn the clock back at this point-- we must go forward."
But the facts on the ground are not the creation of states-- they are the creation of lower federal courts. Opponents of marriage equality tried to create a bandwagon effect in the states against same-sex marriage by piling up state constitutional amendments in the early 2000s. In this way, they hoped to keep a second bandwagon effect in favor of same-sex marriage from ever forming. However, they neglected to account for changes in national popular and elite opinion, which provide an alternative avenue for constitutional change when change at the state level is blocked or preempted. Thus, we now have the simulation of a bandwagon effect in the states created by lower federal courts, and blessed by the United States Supreme Court. You can call this a federalism story if you like, but it seems much more complicated than that.