Leftists have attempted to move the goal posts with regards to incarceration using all kinds of secondary, or irrelevant, criteria for evaluating it's efficacy. Criteria such as "how well does it rehabilitate?", "how well does it deter?", "how much does it cost?" etc. How well it rehabilitates or deters are at best secondary considerations, and if it does neither, but succeeds at reducing crime by simply keeping criminals off the street, then it is very effective at what it is primarily supposed to be effective at. As for cost, statistics citing the enormous cost of incarcerating criminals, which liberals tend to spew at every given opportunity, are actually largely irrelevant because they are rarely, if ever, presented in the context of what incarceration costs versus the cost of having those criminals let loose in society. There is much evidence that suggests that the costs of incarceration are less than those potential costs of releasing thousands of criminals into the streets (or failing to incarcerate them in the first place), and Sowell lists some of those evidences in his book.
Some potent evidence for the efficacy of traditional incarceration practices are the witnessed results of alternative philosophies of incarceration, namely focusing on prevention by getting at "root causes" and seeking to "rehabilitate".
In the United States, where murder rates had been going down for decades, and were in 1961 less than half what they had been in 1933, the legal reforms of the 1960s - applying the ideas of the intellectuals and widely applauded by the intelligentsia - were followed almost immediately by a reverseal of this long downward trend, with the murder rate doubling by 1974. In Britain the ascendancy of the same vision of crime was followed by similarly sudden reversals of previous downward trends in crime rates. As one study noted:
Scholars of criminology have traced a long decline in interpersonal violence since the late Middle Ages until an abrupt and puzzling reversal occurred in the middle of the twentieth century
Liberals often cite recidivism rates as evidence that "incarceration doesn't work." Which is only evidence that incarceration doesn't work that well at reforming criminals, but reforming criminals is not the primary goal of incarceration. The primary goals of incarceration are protecting citizens from criminals and decreasing the amount of criminal activity that takes place in society. Anything else is extracurricular, and often focus on these extracurriculars can detract from the main goals. Sowell writes:
In Britain, as in the United States, it is often taken as axiomatic that "prisons are ineffective," as The Economist put it. The reason: "They may keep offenders off the street, but they fail to discourage them from offending. Two-thirds of ex-prisoners are re-arrested within three years of being released." By this kind of reasoning, food is ineffective as a response to hunger because it is only a matter of time after eating before you get hungry again.
Some more reasonable liberals will object something along the lines of "Well, OK, but aren't some sentences too harsh? Doesn't America incarcerate too many people, and for too long?" Which is a legitimate question to ask, and there certainly is a theoretical threshold at which one can go too far with a policy. But I don't think we've reached that threshold as crime is still too high, and incidents of unjust, excessive incarceration, or of unjust laws, are fairly low. Though, with that being said, I don't necessarily think all of our resources are optimally directed at this time. That is, there's probably too much focus on minor drug offenders, for example. So certainly our system is not perfect. It's just the best.
Liberals will also list anecdotal evidences of 'unjust' sentences for minor crimes, like the third strike offender whose third strike was stealing a pizza. Often concluding that some mandatory sentencing law, such as the three strikes law, is too harsh. Because it can result in people who commit three nonviolent crimes serving very long sentences. I fundamentally disagree with the notion that this is unjust because I think a proven repeat minor criminal has higher potential to be someone who commits a serious crime. Or, minimally, will continue to commit minor crimes costing society a lot to get hurt by them, catch them, judge them, jail them, release them, catch them, judge them, etc. ad infinitum. Everyone deserves a second chance. And a third. But not much more, at least from society (from each other as individuals and from God is another matter). Not to mention that liberal defense lawyers, and their ilk, right after decrying the injustice of a third strike life sentence for a nonviolent offender, will make the same argument on behalf of a rapist or a child molestor. These people's objections should be taken very lightly.
Even if I did recognize some anecdotal examples of unjust sentences, that doesn't even begin to make the case that the entire system of incarceration is unjust. Any functioning justice system will have instances of injustice, injustices done to both criminals and their victims by the way (the latter examples of injustices being conspicuously missing from such lists). The task is minimizing, not eliminating, these injustices. And we do an excellent job of that already.