This is obviously not the main issue driving anyone's thinking about marriage equality, but throwing out the Defense of Marriage Act as the Supreme Court did today should have some meaningful implications for the federal budget. The research that exists on this is a little sketchy and uncertain, but the main conclusion is that it will likely have a small positive impact on deficits. The ur-text here is a 2004 Congressional Budget Office analysis that was done at the behest of Ohio Republican Steve Chabot, who I think was hoping for the opposite conclusion.
The main issues are threefold. Increased government payouts of spousal benefits to public employees and Social Security beneficiaries, lower government payouts of benefits for means-tested programs, and the fact that married people are treated differently for tax purposes. Chabot probably had the first issue in mind when he asked the CBO to look at this, but he forgot the long-standing conservative nostrum that marriage is one of the most effective anti-poverty programs around. But it is! And even though low-income gay and lesbian couples have very little cultural visibility, the Williams Institute at UCLA has recently shown that they very much exist. The CBO estimated that savings from less spending on means-tested entitlements would cancel out the increased cost of spousal benefits, leaving taxes as the decisive factor. The way the federal tax code is structured, married people with similar incomes pay a "marriage penalty" while married people with disparate incomes gain a marriage benefit. CBO estimated that between these two factors you'd end up with about $400 million a year in extra tax revenue.
This is a slightly outdated analysis in two ways. One is that the tax code has changed slightly since 2004 so the revenue boost will probably be a bit bigger than that. More importantly, the Affordable Care Act created a large new means-tested program, so, relative to the new baselines, marriage will cut spending more.
On the state level I have yet to see as rigorous an analysis as the Williams Institute's 2009 look at Maine. They found a positive fiscal impact in which reduced Medicaid payments more than offset increased employee benefit costs. But, of course, Maine is just one small state. Different states' Medicaid programs are structured differently, state income tax policies vary widely, and I assume spousal benefits offered to public employees also vary quite a bit. That said, generous Medicaid programs tend to correlate politically with generous treatment of government employees, so, in general, this should tend to balance out.
In terms of the larger economy, the economic benefits of marriage equality in the short-term are unambiguous. You get a one-off surge in weddings with all the associated economic activity. In the longer term, we just recapitulate the larger debate. Married people generally do better economically, so if you correctly think that marriage equality will bolster American marriage by allowing it to remain a relevant and modern institution for 21st-century life, then you'll see long-term benefits. If you wrongly think that an end to discrimination will lead to the unraveling of heterosexual marriages, then, of course, you'll see a disaster here.