Cap-and-trade programs for air emissions have become the widely accepted, preferred approach to cost-effective pollution reduction. One of the important design questions in a trading program is how to initially distribute the emissions allowances. Under the Acid Rain program created by Title IV of the Clean Air Act, most emissions allowances were distributed to current emitters on the basis of a historic measure of electricity generation in an approach known as grandfathering. Recent proposals have suggested two alternative approaches: allocation according to a formula that is updated over time according to some performance metric in a recent year (the share of electricity generation or something else) and auctioning allowances to the highest bidders. Prior research has shown that the manner in which allowances for carbon dioxide (CO2) are initially distributed can have substantial effects on the social cost of the policy as well as on who wins and who loses as a result of the policy. Another concern with a regional cap-and-trade program like the Regional Greenhouse Gas Initiative (RGGI) is the effect that different approaches to allocating emissions allowances will have on the level of CO2 emissions outside the region, commonly called emissions leakage. In this research we model historic, auction, and updating approaches to allowance allocation that we call bookends, then model various variations on these approaches. We consider changes in measures such as electricity price, the mix of generation technologies, and the emissions of conventional pollutants inside and outside the RGGI region. We examine the social cost of the program, measured as the change in economic surplus, which is the type of measure used in benefit-cost analysis. We also examine the effects of different approaches to distributing allowances on the net present value of generation assets inside and outside the RGGI region. We find that how allowances are allocated has an effect on electricity price, consumption, and the mix of technologies used to generate electricity. Electricity price increases the most with a historic or auction approach. Coal-fired generation in the RGGI region decreases under all approaches but decreases the most under updating. Gas-fired generation decreases under historic and auction approaches but increases substantially under updating. Renewable generation increases under historic and auction approaches but decreases slightly under updating as a consequence of the expanded generation from gas. Consistent with the changes in the composition of generation, the decline in emissions of conventional pollutants including sulfur dioxide (SO2), nitrogen oxides (NOx), and mercury that was expected as a result of the Clean Air Interstate Rule is accelerated substantially as a result of the RGGI policy, particularly under updating. The cost of complying with SO2, NOx, and mercury rules declines similarly. We find that the social costs of the bookend auction and historic approaches are comparable and that the social cost of updating is roughly three times that of the other approaches. At the same time, updating yields greater emissions reductions on a national basis (because it produces less emissions leakage) and greater cumulative reductions in emissions at the national level than historic allocation. Varying the design of the updating approach can reduce its social costs but generally would increase leakage at the same time. An updating approach with allocation to all generators, including all nuclear and renewables has the lowest social cost within the RGGI region of any policy analyzed, although this result comes at the expense of costs imposed outside the region. When the approaches to allocation are mixed, we find the changes in electricity price, generation, and emissions are roughly a combination of the performance of each individual approach. In particular, social costs typically are lower under the scenarios that combine an auction with updating than when updating is the exclusive approach to distributing allowances. Who wins and who loses from the policy varies with the approach to allocation. Under a historic approach, producers in the RGGI region gain substantially and generally are better off than without the program; such is not true under an auction or updating. Producers also gain overall from the policy when a historic allocation is combined with an auction, but the gains are substantially less than in the 100% historic case. Producers outside the region tend to benefit considerably from the higher electricity price in the RGGI region but benefit the least under updating because the effect on electricity price is lowest. Consumers both inside and outside the RGGI region are adversely affected under all allocation approaches but much less so under updating because the change in electricity price is lowest. One exception is when eligibility for allowances under an updating allocation is limited to nonemitters only, in which case the electricity price increases substantially. Different types of generators fare differently under the various allocation approaches. Asset values for all types of generators are highest under a historic approach, although the difference between historic and auction approaches is small for nuclear generators. Compared with the baseline, both nuclear and existing gas-fired generators in the RGGI region gain under an auction. Only gas-fired generators gain under the bookend approach to updating, although nuclear generators benefit as well under updating designs that include them among those eligible for allowances. Coal-fired generators lose the most under updating. Moving from 100% updating to auctioning an increasingly larger share of allowances generally has a positive effect on asset values for all fuel types including coal. The one exception is that moving from 50% auction and 50% updating to 100% auction has a negative effect on the asset values for coal. Finally, we conduct sensitivity analyses with higher natural gas prices and constraints on electricity transmission capability. The social cost of the RGGI program does not appear to be sensitive to these constraints. Higher gas prices or transmission constraints alone impose significant costs that are larger than the effect of adding the RGGI policy. For example, their substantial effect on electricity price is greater than the added effect imposed by the RGGI program. The constraints that are modeled do not appear to have a strong impact on RGGI implementation. We also conduct a sensitivity analysis with renewables portfolio standard policies in place throughout the region. The resulting prices of electricity and CO2 emissions allowances are slightly lower than without the renewables policy.