The objective was to reduce the incidence of liver abscesses (LAs) in young bulls by reducing the starch content of the concentrate and increasing the straw intake by adding molasses without reducing performance. Eighty-five Danish Friesian bulls (146 ± 2.6 kg live weight (LW)) were allocated to four treatment groups in a 2 × 2 factorial design and were ad libitum fed either a high (H) (43%) or a low (L) (25%) starch concentrate, and either chopped barley straw (S) or a mixture (Sm) of chopped barley straw and sugar-beet molasses (75:25). The bulls were slaughtered at 440 (±4.2) kg LW (11.3 months of age) on average. The concentrate dry-matter intake (DMI) was 7.5% higher (P < 0.001), the roughage DMI was 12% higher (P < 0.01), whereas the total net energy intake was 7% lower (P < 0.001) for the L compared with the H concentrate. By adding molasses to the straw, roughage DMI increased from 0.50 to 0.96 kg/day (P < 0.001). Average daily gain (ADG) was not affected by treatment. Feed conversion efficiency (FCE) for dry matter (FCEDM) was decreased (P < 0.001), whereas FCE for net energy (FCENE) was increased with the L concentrate (P < 0.01), suggesting a lower utilisation of the cell wall fraction with the H starch concentrate. There were no major effects of treatment on carcass characteristics. The higher roughage intake with Sm prevented rumen wall damage, but did not reduce the number of animals with LAs. The L concentrate did not affect the rumen wall but reduced the level of LA (2 v. 9 animals, P < 0.02). There was a higher level of respiratory diseases in animals fed the H concentrate (P < 0.05). The experiment showed that it was possible to maintain performance and reduce LA by using a lower starch content of the concentrate.