Pablo Escobar was once worth $9 billion.
Carlos Lehder was worth up to $2.5 billion.
These numbers sound like liquid net worths, because cartels usually don't release their numbers, and nobody's quite sure what the P/E ratio ought to be. Illegal drugs are a product with: 1) a very high markup, 2) a tendency to be produced by monopolies, and 3) very high risk of arrest or murder for sellers. It seems to me that this guarantees that a few people in the drug trade will simply have more money than anyone could possibly spend, and they won't pay taxes on it.
Consider Manhattan in 1985, borrowing liberally from Brett Easton Ellis. Let's say there are 10,000 bond traders with a $100/day coke habit. And let's say the profit margin on cocaine is a very conservative 90%. Each year, they transfer $328,000,000 to Mr. Escobar and his associates. Now, any individual bond trader could find a way to blow his $36,500/year on something besides, well, blow. But Mr. Escobar would be hard-pressed to spend more than 1% of his money on luxuries, so he has to save it. And what are the effects of that?
Either a) he invests in more coke-producing capacity (but only if that raises net profits, which just means deferring the question), b) he holds on to cash, which keeps bills out of circulation and fractionally lowers the money supply, making net lenders richer and net borrowers poorer, or c) he invests that money, turning consumption in Manhattan into capital everywhere else.
The answer is clear: if the government wants to increase our savings rate, the first thing they need to do is increase the number of Masters of the Universe (unfortunately, we may already be at capacity). The second thing the government needs to do is more stringently enforce laws against using cocaine, without trying to lower the demand (seems we're already doing that, too). And finally, if none of that works, the cheapest way to increase global savings rates is to subsidize the next Pablo Escobar.