Read the ten-ramen-pack opinion carefully and you do not find malice. You find a judge who has, internally, understood the case as a re-application of established doctrine. The defendant has prior petty-theft records. Habitual theft (Article 332) carries an aggravation. The recommended range, with that aggravation, is in actual-imprisonment territory. The judge does not feel free to invent mitigators that the law does not recognize. So the judge applies the rule.
That is the steel-man for the conservative side of the legal-stability vs. concrete-justice debate. If judges felt entirely free to invent mitigators, the rich would always have invented mitigators waiting for them, and the poor would have judges' moods. Predictable rules protect citizens from arbitrariness. That argument has weight, and it has weight especially for citizens who would otherwise be at the mercy of judicial whim.
The complication, as the data in our database shows, is that mitigators are not invented — they are inherited. The Sentencing Commission's published guidelines list nameable mitigators. 'Sincere efforts at restitution' is one. 'Social contribution' is one. 'Voluntary disclosure' is one. These are not new inventions for chaebol cases; they are the standard menu. What differs across our pairs is which defendants can credibly claim them.
The job-seeker cannot claim 'social contribution' because we do not yet have a doctrine in which a hungry person trying to keep themselves alive is a contributor to society. We could have such a doctrine. Other legal systems have produced versions of it under headings like 'extreme necessity' or 'subsistence defense.' Korea's law does not currently centralize one. The result is that the same mitigator menu, with the same items on it, produces structurally different outcomes — not because the judges are different, but because the menu was written when defendants like the chairman were expected and defendants like the job-seeker were not.
This is the conservative-liberal pivot for sentencing reform. The conservative position is correct that judges should not invent doctrines. The liberal position is correct that the existing doctrines have a thumb on the scale. The reform path that takes both seriously is to expand the menu: to recognize 'subsistence circumstances' or 'desperate restitution' as standard mitigators, eligible for credible claim by defendants who currently cannot access the menu. That change would not require judges to be more 'merciful.' It would require the menu to be more honest about who actually cooks from it.