BROADCASTS.com

b'

Here\\u2019s a graph of US air pollution over time:

\\xa0

During the discussion of\\xa090s environmentalism, some people pointed out that this showed the Clean Air Act didn\\u2019t matter. The trend is the same before the Act as after it.
This kind of argument is common. For example, here\\u2019s the libertarian Mercatus Institute\\xa0arguing that OSHA didn\\u2019t help workplace safety:

\\xa0

I\\u2019ve always taken these arguments pretty seriously. But recently I\\u2019ve gotten more cautious. Here\\u2019s a graph of Moore\\u2019s Law, the \\u201crule\\u201d that transistor counts will always increase by a certain amount per year:

\\xa0

The Moore\\u2019s Law Wikipedia article lists factors that have helped transistors keep shrinking during that time, for example \\u201cthe invention of deep UV excimer laser photolithography\\u201d in 1980. But if we wanted to be really harsh, we could make a graph like this:

\\xa0

But the same argument that disproves the importance of photolithography disproves the importance of anything else. We\\u2019d have to retreat to a thousand-coin-flips model where each factor is so small that it happening or not happening at any given time doesn\\u2019t change the graph in a visible way.
The only satisfying counterargument I\\u2019ve heard to this is that Moore\\u2019s Law comes from a combination of physical law and human commitment. Physical law is consistent with transistors shrinking this quickly. But having noticed this, humans (like the leadership of Intel) commit to achieve it. That commitment functions kind of as a control system. If there\\u2019s a big advance in one area, they can relax a little bit in other areas. If there\\u2019s a problem in one area, they\\u2019ll pour more resources into it until there stops being a problem. One can imagine an event big enough to break the control system \\u2013 a single unexpected discovery that cuts sizes by a factor of 1000 all on its own, or a quirk of physical law that makes it impossible to fit more transistors on a chip without inventing an entirely new scientific paradigm. But in fact there was no event big enough to break the control system during this period, so the system kept working.
But then we have to wonder whether other things like clean air are control systems too.
That is, suppose that as the economy improves and stuff, the American people demand cleaner air. They will only be happy if the air is at least 2% cleaner each year than the year before. If one year the air is 10% cleaner than the year before, environmentalist groups get bored and wander off, and there\\u2019s no more progress for the next five years. But if one year the air is only 1% cleaner, newly-energized environmentalist voters threaten to vote out all the incumbents who contributed to the problem, and politicians pass some emergency measure to make it go down another 1%. So absent some event strong enough to overwhelm the system, air pollution will always go down 2% per year. But that doesn\\u2019t mean the Clean Air Act didn\\u2019t change things! The Clean Air Act was part of the toolkit that the control system used to keep the decline at 2%. If the Clean Air Act had never happened, the control system would have figured out some other way to keep air pollution low, but that doesn\\u2019t mean the Clean Air Act didn\\u2019t matter. Just that it mattered exactly as much as whatever it would have been replaced with.