But it's not true as a general proposition, nor are you excluding an irrelevant class of litigation. Much (most?) commercial litigation is between well-funded parties (e.g. a big company and it's insurer).
> Interesting report, thank you for the link. I think 90% in and of itself is high enough to warrant scrutiny
The government should only bring a lawsuit if it looks at the evidence and decides it can prove guilt beyond a reasonable doubt. Generally, that's in the 95% certainty range. A jury should only convict if it comes to the same conclusion. Thus, it should be quite unusual for the jury to acquit. If the jury is acquitting in 25% of cases, that means the government is bringing lots of prosecutions it shouldn't be bringing.
It's true that over the last 30 years, the government has gotten much more conservative about bringing prosecutions. That's a good thing. Many might disagree with the law underlying, say, the Silk Road case, but it's hard to say the government didn't base that prosecution on a wealth of evidence.
> To the contrary, 90% is about right. The government should only be bringing lawsuits if it has a good faith belief it can establish the facts beyond a reasonable doubt. BRD is about 90-95% probability.
There is a huge fallacy underlying your numbers. 90% is about right if the number of false convictions is acceptably low. 90% is a terrible number if it should contain a large number of false convictions.
The 'establishing of facts beyond reasonable doubt' only happens for cases that actually go to court, in the light of the above mentioned asymmetry in resources you're going to end up with a lot of cases that will never make it that far. Some will claim this is efficiency but I don't really feel comfortable with an 'efficient' process that places large numbers of people in lock-up. The chances of getting it wrong are fairly high in a system that aims for a certain percentage of convictions, after all there are two solutions to getting that percentage, the first is to ensure that only the 'good' cases make it through the system, the other is to railroad the bad ones through as well and make sure that a conviction is secured anyway.
In fact, one could argue the only way to look at it is in a precision/recall manner, how many cases that were known to be solid turned out to be false positives and how many cases that were dropped were the false negatives (cases that should have been won but were not, for instance because the defendant was wealthy enough that they could afford a large defense).
This is of course all only relevant if you actually care about justice rather than about convicting as large a number of people possible once they enter the judicial system, and I'm not at all sure that that is the goal of the systems in place in the various countries. Too many people depend for their livelihood on large numbers of people getting convicted.
> 90% is about right if the number of false convictions is acceptably low. 90% is a terrible number if it should contain a large number of false convictions.
Sure. But you implied above that a 90% conviction rate warrants more scrutiny than a 75% conviction rate. To the contrary, a 90% conviction rate may be indicative of a fair process. A 75% conviction rate is definitely indicative of an unfair process.
Obviously you still need to scrutinize the rate of false convictions. But that's true regardless of the conviction rate. Indeed, a low conviction rate suggests that the government is bringing marginal cases and trying to make things "stick" which increases the risk of false convictions.
> Some will claim this is efficiency but I don't really feel comfortable with an 'efficient' process that places large numbers of people in lock-up.
The U.S. has a large number of people in prison because it imposes long prison sentences for some very common and easy-to-prove things: drug sales, auto theft, assault, firearms. If you take a random sampling of federal convictions, you'll see that almost always there is a laundry list of physical evidence of guilt. The government doesn't bring marginal cases.
The solution to the prison problem is not to force all these clear-cut cases to go to trial. It's to stop imposing 20-year prison sentences for people selling a few pills of Oxycontin, or stop putting felons back in prison for 10 years just for having a gun.
>It's true that over the last 30 years, the government has gotten much more conservative about bringing prosecutions.
I'm not sure that this is actually true. Instead, I see a much greater use of plea bargains, especially of the "accept this deal or we will throw every possible charge at you" type. Given the sheer number of laws, this can lead to life-ending penalties for seemingly minor crimes.
But it's not true as a general proposition, nor are you excluding an irrelevant class of litigation. Much (most?) commercial litigation is between well-funded parties (e.g. a big company and it's insurer).
> Interesting report, thank you for the link. I think 90% in and of itself is high enough to warrant scrutiny
The government should only bring a lawsuit if it looks at the evidence and decides it can prove guilt beyond a reasonable doubt. Generally, that's in the 95% certainty range. A jury should only convict if it comes to the same conclusion. Thus, it should be quite unusual for the jury to acquit. If the jury is acquitting in 25% of cases, that means the government is bringing lots of prosecutions it shouldn't be bringing.
It's true that over the last 30 years, the government has gotten much more conservative about bringing prosecutions. That's a good thing. Many might disagree with the law underlying, say, the Silk Road case, but it's hard to say the government didn't base that prosecution on a wealth of evidence.