There is no case for treating young offenders differently. The emphasis has to be on responsibility. You have to concentrate on locking people up in the most human and civilized way. Young criminals are cultural criminals, a lot smarter than the treaters. You cannot deal with social injustice through the criminal justice system. If you try to correct in this way, you corrupt the system.
Persistent young offenders are not beyond help and special efforts should be made to rescue them from a life of crime. This will involve taking them from their home environment and friends, but this environment has often contributed to the criminality. Secure training units (juvenile detention centres) makes it clear to young offenders that the public will no longer tolerate their anti-social behaviour. Opposition to these proposals does not appear to come from the most vulnerable victims, but from certain sections of society that are to some extent insulated from persistent juvenile criminals, and that may be reflected in their views.
UK statistics show all previous forms of juvenile incarceration have failed. Of juveniles sentenced to community punishments (without incarceration), about half will re-offend; the figure for those released from custody is at least 82%. Approved schools were discontinued 20 years ago because their alumni reconviction rate was 49% higher than juveniles treated in other ways. The point is that child criminals are still children. There is a wealth of research to show that even "really persistent" juvenile offenders mostly grow out of crime; treat them wrongly, and you may criminalise them for life, leading the regular car thief to burglary, armed robbery or worse.
"Persistent" offenders cannot be differentiated from other children with profound social needs. Of the 531 persistent offenders (arrested 3 times or more; 1992, UK), 43% were "known" to social services and 2.5% were on the child protection register at the time of their arrest. Research in 3 UK cities showed that incarcerating the 20 most persistent offenders in each city would reduce the overall reported crime rate by less than 2% in the first 12 months. Thereafter, if those individuals are released, with predictable rates of re-offending, overall crime is predicted to rise. Young offenders are a threat to victims, they are disruptive to social life, they are at times a dangerous nuisance, but that does not justify millions of public expenditure on a flawed policy of increased custody which is already a predictable failure.
Locking children up is a Dickensian notion that focuses on punishment. These kids are causing trouble because they are in trouble. Their intention is not to commit a crime against the community, but to send a message that they need help. There are no evil children, just evil actions. Locking them up is the lazy way and the cheap way. Proper family care and therapy costs a fortune.