Reinforced concrete structures are subjected to several sources of deterioration that can cause a progressive reduction in performance. Among the existing infrastructure typologies, reinforced concrete half-joint bridges are particularly critical. This common structural configuration is characterised by a reduction in depth at the supports. The underside of the full-depth section does not benefit from support confinement in the anchorage zone of the longitudinal tensile reinforcement. Moreover, leakage of contaminated water often causes corrosion of the internal steel reinforcement and cracking of the concrete cover. Thus, half-joints are particularly vulnerable to bond degradation due to corrosion and cracking. To develop an improved understanding of deterioration in reinforced concrete, and half-joints in particular, three experimental programmes were carried out. The first investigation consisted of accelerated corrosion and concentric pull-out testing on unconfined concrete specimens. It was found that bond degradation only occurred where corrosion products caused expansive pressure, splitting cracks and a progressive loss of rib interlock. The reduction in bond depended upon the widths of the cracks, irrespective of the levels of corrosion. Surface crack widths resulted in better indicators of corrosion-induced bond degradation than conventional measures of corrosion such as mass loss or attack penetration. In the second investigation, a novel bond test set-up was developed to remove the parasitic effect of external confinement from the support reactions. The new geometry was adopted to study the effects of internal confinement on bond. The results showed that the dependence of anchorage capacity upon the concrete cover distance was not monotonic in the presence of confining reinforcement. Within a range of optimum cover-to-diameter ratios, small splitting cracks activated the transverse reinforcement and led to a bond strength uplift. In the third investigation, half-joint beams were tested in three-point bending to study the consequences of anchorage degradation on the overall behaviour of the component. Analytical predictions based on the lower bound theorem of plasticity were found to be overly conservative and did not capture the behaviour observed experimentally. This was attributed to a combination of local effects of greater anchorage capacity than predicted and global effects relating to alternative load-paths not taken into account by the predictive models. It was concluded that local deterioration led to a premature local failure of the half-joints, whilst the rest of the structural component was far from its ultimate behaviour and little redistribution had occurred. These conditions undermine the assumptions that underpin plasticity-based models for new design, thus reducing their accuracy when used for the assessment of existing structures in the presence of deterioration. The findings of this research contribute to the development of an effective assessment approach that correlates visual and measurable parameters on the outer surface of the concrete (e.g. crack widths), deterioration processes hidden inside the structure (e.g. bond degradation) and their structural consequences (e.g. overall strength reduction). Improved assessments could lead to enhanced asset management strategies that would reduce the safety risks, maintenance costs and environmental footprint of the infrastructure network.