New Zealand and Tort Law; Arbitration

Over morning coffee I read Peter Schuck on tort law in New Zealand (hat-tip and link, Larry Solum). New Zealand has a no-fault system for personal injuries generally, and something similar for malpractice. Processing costs per case are of course low, but the system may illustrate a perverse effect: low transaction costs per case can lead to a high total amount of transaction costs because more claims are made. The paper doesn't compare total transaction costs per capita in the US and New Zealand, but it says "In 2006-07, the ACC had pending 1.6 million claims, ... this for a population of 4.3 million..."

I wonder what the total transaction costs are? Probably in New Zealand many claims are made that are either harder to prove fault in or too small to be worth hiring a lawyer in the US. Actually, we also need to figure in the insurance company transaction costs, since both in the US and New Zealand private insurance companies exist, although New Zealand has a much bigger free public medical care system than the US. (Another complication is that US medical care, "gold-plated", is more expensive.) Anyway, is it good or bad that lesser claims get compensated in New Zealand? Is it better to just let losses lie for them, and avoid transaction costs altogether?

There are many ideas for tort reform. I wonder if any country uses government arbitration? Here's my idea. After an accident, a policeman must be called in if there is to be compensation to anyone (the hospital or morgue will do it if the injury is severe enough). He writes a report and interviews witnesses. Each side then writes a report saying, under oath, what they claim happened. A government arbitrator looks at the reports, maybe interviews witnesses (without lawyers present), and assigns the case to one of three categories: X wins, Y wins, or Unclear. The government pays claims for Unclear cases; private insurance or the people themselves for others. Insurance companies are free to use the results for experience rating.