The current U.S. wetland mitigation policy of "no net loss" requires that a new wetland be created to replace any natural wetland destroyed under development pressures. This policy, however, may be resulting in a net loss of carbon-based wetland functions. We evaluated the ability of created wetlands to accumulate carbon and to mitigate loss of carbon-based functions in natural wetlands with variable hydrology. Potential limiting factors to carbon accumulation within created systems included soil aggregation, texture, and bulk density. Rates of soil development and the time required for created wetlands to accumulate the amount of carbon found in natural wetlands were estimated by an exponential model. Soils collected from five created (ages 3-8 years) and four natural freshwater marshes, located in central Ohio, USA, were analyzed for soil organic carbon (SOC), mineralizable soil carbon (Cmin), water-stable aggregates (WSA), particle-size fractions (PSD), and bulk density. Peak-standing aboveground plant biomass was also quantified. Created wetlands contained significantly less plant biomass, SOC, and Cmin than natural wetlands (c < 0.05; false discovery rate). Soil physical properties also differed significantly between created and natural wetlands, with fewer macroaggregates, more microaggregates, more silt-clay (0-5 cm only), and higher bulk density in created wetlands (a < 0.05; false discovery rate). Carbon content was positively correlated with macroaggregate content and negatively correlated with microaggregate content, silt-clay fraction, and bulk density. Fit of SOC data to the exponential model indicated that a newly created wetland would require 300 years to sequester the amount of SOC contained in a natural wetland. At this rate of carbon accumulation, a mitigation ratio of 2.7:1 (area) would be necessary for successful mitigation over a 50-year time period. However, other trajectories fit the data equally well and suggested area mitigation ratios of 2.2:1 (logistic) to 4.4:1 (linear regression) to 5.1:1 (exponential regression). Whether created wetlands are on a trajectory toward natural wetland carbon function, however, remains uncertain. Until gaps in the data are filled and a trajectory verified, the best mitigation policy will be a conservative one, with a restrictive permitting process and high mitigation ratios (5.1:1 minimum).