Using this calculator, pick a zip code in the Peninsula, I put in a 4000 watt system, estimate my bill to be $200/month, assume 2.5% increase in electricity prices per year, and assume I pay in cash. This gives me a 15 year payback period. At $100/month, it's a 24 year payback period.
If so, now consider that your $200/month electricity bill is the equivalent of $124.24 x $200 or $24,847 worth of "house".
Our actual bills were $250 - $300/month pre-panels and effectively about $15/month post panels. In terms of money returned by 7 years we had paid less in electricity than we had paid for the panels. But that discounts what $20,000 would have been worth if we had invested it (that is the 'do nothing' alternative since you aren't really considering putting that money under a mattress :-) And of course we also made our house more efficient which we would have gotten that pay back either way.
The only downside to our system has been trying to explain to people why it doesn't provide power when the power is off.
[1] http://hfa3741.hubpages.com/hub/How-Much-Does-a-200-000-Hous...
First I think most people are paying closer to $5 per Watt, and I know someone who just refinanced his house in Long Beach at 3.5%, fixed. For comparison, the calculator assumes $10 per watt, and 8%. If you use halve the installed cost, there is obviously a big effect on the payback period (down to 6.5 years with tax incentives).
By calculating my what my electricity bill would have been vs what it was, they have saved me more than their cost (at some point I should go back again and get exact numbers).
Its entirely unclear how much "additional" value they give the resale value of my house. It is clearly some, but I don't have a good way of knowing if it is $20,000 or more.
At the time we financed the purchase as part of a refinancing of the house, so we did know the interest cost for the money, we also went through the great Recession so in traditional financial analysis (the 'do nothing' option) we would have predicted our money would have had a better return than it actually did, although that didn't really change the math all that much.
So I can't see how you don't get it paid off in < 10 years these days. At this point I'm thinking of adding another string to offset the addition of an electric car.