FORTUNE -- Long before he sentenced Rajat Gupta to prison, U.S. District Judge Jed Rakoff had a celebrated and controversial career on the bench. His judicial opinions were thoughtful, direct, and witty. (And, on occasion, reversed by higher courts.) In 2009 he rebuked the SEC for what he found to be insufficient punishment of Bank of America for nondisclosure violations. In 2002 he -- unsuccessfully -- declared the federal death penalty unconstitutional.
The past year was particularly remarkable for Rakoff. In May 2012 he presided over the $163 million settlement between the owners of the New York Mets and the trustee for the victims of Bernard Madoff's Ponzi scheme. And then, in late October, Rakoff sentenced a once highly respected business executive -- Gupta, a Goldman Sachs (GS) director who used to run McKinsey & Co. -- to two years in prison, plus fined him $5 million, for insider trading. Federal sentencing guidelines, which are not mandatory, called for roughly four times that; the government had recommended even more jail time.
Gupta, 63, was convicted of leaking boardroom secrets to Raj Rajaratnam, the billionaire founder of the Galleon Group hedge fund. (Rajaratnam himself was convicted of securities fraud in 2011 and is serving an 11-year term.) Rakoff's sentence for Gupta and the 15-page written opinion that went with it have been cited both as a model of fairness and as an example of undue leniency for white-collar miscreants. "The fundamental problem of this sentence," Rakoff wrote, is that "Gupta's history and characteristics starkly contrast with the nature and circumstances of his crimes."
Sitting judges rarely give interviews. But the 69-year-old Rakoff, a former prosecutor and white-collar defense lawyer who was appointed by President Clinton in 1996, sat down recently with Fortune in his chambers in the U.S. courthouse in lower Manhattan. While Rakoff would not discuss the Gupta case because an appeal is pending, he made clear his objection to inflexible sentencing guidelines. In a wide-ranging conversation, Rakoff also talked about the public's revulsion with white-collar crime and the paucity of prosecutions after the financial crisis, as well as the quality of mercy. Edited excerpts:
GOOD PEOPLE, EVIL DEEDS
Q: I understand we can't discuss Rajat Gupta, but let me ask you a hypothetical. Consider a white-collar defendant who hasn't led a particularly exemplary life. He's just a rapacious capitalist. If I've read your sentencing opinions correctly, that defendant would be in a worse sentencing position than a defendant convicted of the same crime who had done good deeds?
A: Yes, for two reasons. One is that under federal law I'm required to take account of the personal history and character of the defendant. But also because of what I think is an appropriate way to look at sentencing, which is, there are defendants who are good people who have nevertheless done some bad things. I'm not interested at all in people who make charitable contributions, but I am interested in where people have spent their time doing good.
There are people you might describe as evil, to use an old-fashioned term. And I've given white-collar defendants 20 years, like Marc Dreier [a Manhattan lawyer convicted in 2009 of a scheme to defraud investors of $700 million]. I found very little that was redeeming in his character, and he committed a huge fraud.
The Gupta sentence seems clearly to be more about what is on the "plus" side of the ledger.
That's a reasonable inference. But that's not all there is to sentencing. It's interesting how the American public, I'm afraid, is very punitive. We have more people in prison on a per-capita basis than any country in the world by a substantial margin: well over 2 million. And when someone does something wrong, something deep in the American character says, "Send him away forever."
That probably explains our historic love for the death penalty.
I think that's right There is a very strong moral streak in American culture. That can be good. I haven't been shy in my own opinions about expressing moral outrage when I thought it was called for. But there is a danger of that being transmuted into a kind of lynch-mob psychology. And when I do a light sentence, I always get some nasty mail along the lines of "You fool. Why did [I] play by the rules? Go out and commit a crime and all you'll get is a slap on the wrist?"
AN ADVANTAGE FOR WHITE-COLLAR CRIMINALS?
Let's compare two people. One is the white-collar criminal who steals and, by virtue of his job, has the opportunity to be involved in prominent causes. The other is the stickup guy at 7-Eleven. Structurally, doesn't that put the white-collar criminal in a better position in a sentencing regime that takes into account one's goodness? It's a flawed analogy.
The kind of person who receives my sympathy is the kind of person who has shown by his acts that he cares about his fellow man and who has acted honorably for maybe 99% of his life and then makes a mistake. I will use Mr. Dreier as a different example. He committed a huge fraud over many years. Yes, he did contribute to charities, but that wasn't the nature of the man. The analogy with the stickup guy would be a guy who led a law-abiding life, but then was desperate and committed a single stickup. I assure you that in those cases, usually in state court, [defendants] are often given very low sentences.
And of course white-collar crime doesn't involve violence.
That's another problem with your analogy. The people who think there's no difference between a violent crime and a nonviolent crime are people, I think, who never experienced a violent crime. Civilization really breaks down in a fundamental way if violence is not dealt with severely. What is interesting is that the public revulsion against white-collar crime over the last 10, 15 years has led to ever-escalating [federal] penalties for white-collar criminals. A sentence is often higher for a white-collar criminal than it is for a violent non-white-collar criminal.
So, white-collar sentences really have changed?
It was unquestionably true 30 years ago that white-collar offenders got lower sentences than drug offenders. That is no longer true, especially after Enron and WorldCom.
What cries out in your sentencing opinions is that this is a really hard process for trial judges.
The kind of sentencing law I like is what's in [federal] Section 3553(a) of Title 18, which basically says, "Here are the factors you've got to look at. We leave it to you to perform this very difficult job of getting sufficiently into the nitty-gritty of how these factors play out" As opposed to this number-crunching nonsense that are the guidelines, which isolate a very few factors and say, Oh, that's going to determine what the sentencing is."
But what happens if 10 smart, reasonable federal judges looking at the same circumstances make very different sentencing decisions?
That of course was one of the reasons for the guidelines, but I don't think that's the cure. The cure, in my view, is robust appellate review to deal with the outliers. But instead they said, "We will try to isolate a few factors -- things that can be measured, like the amount of money lost or the weight of the drugs. We will elevate them above all other factors." And because the guidelines originally were mandatory, it was crazy.
The sociology of sentencing always talks about deterrent value and rehabilitation -- which are vague concepts. Isn't Section 3553's goal of "just punishment" especially so?
There is a moral component to sentencing. Deep in the psyche of human beings, I think, there is a feeling that the world is not right and justice is not being done if bad people aren't punished for bad things. That's what's meant by just punishment.
PROSECUTIONS AND THE FINANCIAL CRISIS
Let's talk about the prosecutions following the financial meltdown. There's been a lot of criticism that few individuals have been thrown in jail other than people like Raj Rajaratnam. Nobody at AIG (AIG), nobody at Lehman Brothers.
I've written about it in several opinions. So for example, in my Bank of America decision where I initially turned down a proposed settlement between the SEC and Bank of America (BAC), I pointed out among many other flaws the fact that no individual was named for what the SEC asserted was a blatant fraud orchestrated from the very top.
You were lauded as a hero for that.
You'll get no argument from me on whether enforcement perhaps has not done as much as it could in the white-collar area. The very different question is: Once someone has been successfully prosecuted, what's the appropriate sentence?
From the public's perspective, aren't the two questions related?
It's very important that justice be equal, and that means you should not be able to escape criminal liability because you're at the top of a major financial organization or because there are too many complexities involved. No one would argue if you were talking about complicated conspiracies that the federal government isn't expert in starting at the low end and flipping people all the way up to the top. And that often has been true in white-collar situations, but seemingly less with reference to the recent crisis.
So why haven't there been more prosecutions?
I was chief of securities-fraud prosecutions back in the 1970s [for the U.S. Attorney's Office in Manhattan]. There are typically two problems in prosecuting white-collar cases. The first is because they turn to a large extent on gathering pertinent information; you either have, as with Raj Rajaratnam, a wiretap, or you have someone you flip who agrees to cooperate, like in Boesky and Milken kind of prosecutions.
Flipping people also gets them a reduction in sentence.
A judge in Russia once asked me, "How can you use those cooperators? These are people who have committed serious crimes, and yet you allow them to escape virtually all jail time. That's immoral." And I said, "Well, how do you go after complicated cases?" And the judge said, "Easy. Unlimited wiretaps." So you pick your poison.
What's the second problem?
All crimes, not just white-collar, require what lawyers call mens rea -- some kind of evil intent. And our system rightly distinguishes between people who do things knowing they're wrong and people who make mistakes negligently. So that's a problem, particularly in complicated financial situations.
It's a value judgment, though. Some negligent things get criminalized.
The typical thing is "reckless disregard." You have the guy who is gunning his car at 100 miles an hour. He's not intending to kill anyone, but, boy, he damn well knows he's increased the risk of doing it. Historically, the concept of reckless disregard has been important in prosecuting high-level white-collar crimes. It's not for me to comment on what the Justice Department or anyone else is doing because I don't know the facts. But I am surprised there hasn't been more use of the reckless-disregard concept in going after situations where you're never going to find a guy who says, "Yes, I decided today to cheat a million people and, Dear Diary, wasn't it a great day."
Is the failure to go after people a problem of prosecutorial will?
I'm sure it's not a lack of will because prosecutors typically make their reputation by bringing great cases. Those cases tend to be very difficult to make. They require a lot of commitment. When I was chief of the fraud unit, I would say to a [subordinate], "Here's a situation that has a bad smell, so I want you to look into it, and we have some agents who are going to help. But I have to tell you upfront that in the end you may spend a year on this and conclude there's nothing that can be prosecuted. And I certainly don't ever want you to be in the position of saying, We ought to prosecute because I spent a year on it.'"
So, for all we know, those yearlong investigations may have taken place in 2009 or 2010?
It may be. I've wondered a bit whether there are as much resources [at the Justice Department] committed to those cases as there should be. I also wonder whether the fact the government itself was involved in some of these situations has made them more difficult to pursue -- the mortgage-based cases and so forth. The government got involved early on in promoting mortgages for everyone. I think it makes [a prosecutor's] job more difficult because it's one thing to use your great subpoena power to bring in an executive to the grand jury. It's quite another thing to say, "And now, Mr. Secretary of X, or subcabinet officer, I want you to come down and testify."
THE SYMBOLIC VALUE OF PUNISHMENT
Once we do prosecute, isn't there something to he said for making an example of folks at the top?
There is a tendency, not wholly irrational, to say, "Well, because he is at the pinnacle of business, and because these are difficult cases to make, we ought to string him up." China carries this to the extreme, with the death penalty for 66 crimes, including grand theft. And if I remember correctly, they impose the death penalty on something like 10,000 people each year. And they justify this on the grounds of "Well, there's no better deterrent than the death penalty." If deterrence is the whole ball game, we should have the death penalty for anyone who exceeds the speed limit by 10 miles an hour. I don't think you could find anyone who would think that fair.
Let me ask it differently: There's a generalized notion about justice that starts with "from each according to his ability ..." So, shouldn't we throw the book at individuals in special positions of trust -- CEOs, university presidents?
It has only a very modest germ of truth. Most people who commit crimes know damn well they're doing wrong, so don't tell me it's only the guy at the top who knows. Lots of white-collar crimes are committed by the bookkeeper who embezzles $10,000.
The argument would he that as you go up the continuum of white-collar crime, you punish people more severely because that's where society needs to have the point reinforced.
It sounds like a little bit of a violation of the 14th Amendment's prohibition against unequal treatment under the law.
We would stop just short of that, Your Honor! The idea is to enforce the policy judgment behind "just punishment."
Here's what I think makes more sense. Everyone agrees a recidivist should be punished more. If a person committed a humongous crime, obviously that's a relevant factor. If the person abused a position of trust, that's relevant. There are a host of factors.
You mentioned the letters you get after a controversial sentencing --
And telephone calls!
You actually take the calls?
When there's no one else around, I pick up the phone, sure. I had the situation where a guy who was clearly a little drunk said, "I saw what you did in the X case, and that was absolutely outrageous, and I'm going to kill you." And I realized the case wasn't mine. It was a colleague's. And I was so tempted to say, "Let me transfer your call."
This story is from the February 4, 2013 issue of Fortune.