Recent reviews of salmonid habitat restoration programs have recommended that managers emphasize strategies that restore natural habitat-forming processes, such as restoring riparian vegetation, over placement of instream structures. In addition to the direct benefits of shading and providing a source for large woody debris (LWD), riparian restoration is often implemented to improve channel morphology for purposes of restoring fish habitat. However, multiple studies provide equivocal evidence that restored vegetation can lead to improved channel form within a period of years to decades. In this study, we examined the effectiveness of riparian restoration for improving channel morphology and fish habitat in four hardwood-dominated streams in Mendocino County, California. These streams support populations of steelhead Oncorhynchus mykiss and contain reaches with riparian corridors that were restored through exclusionary fencing implemented 10-20 years earlier. We compared channel morphology, LWD, and late-summer water temperature between the restored exclosure reaches and geomorphically similar control reaches within the same properties. Channels within exclosures were significantly narrower and had greater heterogeneity in long profile elevation than control reaches. Frequencies of LWD and debris jams were considerably greater in exclosure reaches than control reaches and were comparable to values from similar streams with mature forests. Late-summer water temperature in exclosures was within the acceptable range for steelhead, whereas water temperature in control reaches was warmer and potentially detrimental to steelhead. Riparian restoration in exclosures has resulted in quantitatively improved habitat characteristics and qualitatively different channel morphologies as compared with control reaches. The ability of vegetation to improve channel morphology in this region is probably due to frequent overbank flooding and high sediment loads. Through a comparative analysis of the cost and performance of exclusionary fencing versus those of instream structures, we propose that riparian restoration can produce instream salmonid habitat benefits that are more comprehensive, sustainable, and cost-effective than the benefits generated by instream structures.