I've been away from PG&E-land for too long to know if this still works, but:
1) With solar PV, you generate high-value (high priced) electricity (peak generation is ~10 AM - 4 PM, same as peak sunburn hours
)
2) And, with TOU metering (and especially if you install batteries for a UPS, although that raises installation and operating costs) you can draw a bunch of low-cost electricity at night and then use only battery and solar PV power during the day
which can combine to pay off a solar PV system pretty rapidly, given the steep difference between "peak price" and "off-peak".
Let's say that between refrigerator, air conditioner, hair dryers, and any and all other "power eating" devices (pre-EV), you normally run 1500 kWh/mo with roughly 2/3 of the usage during the day/on-peak and 1/3 at night/off-peak (this is close to the overall state power profile so it's a good guesstimate start point). So, you use 1000 "peak" and 500 "off-peak" kWh.
If you can charge all this at "TOU EV" rate and pay (say) .35 and .10 respectively, that's $350/mo peak and $50/mo off-peak.
To solar-PV-generate 1000 kWh/mo you need about 33 kWh/day. Conservatively (?) estimating sun hours at 4.5, let's call that an 8kW system. At $4/watt installed (that's a low end installed price but not un-achievable), 8kW will cost $32k. If you can take the 30% tax credit, that shaves off $9.6k, leaving $22.4k. At $5/watt installed (more reasonable) it's $40k, minus tax credit of $12k = $28k.
Now, that won't actually eliminate all on-peak usage (generation and usage hours won't always match up) but let's keep running with these. Assume it does take out all on-peak, or $350/mo = $4.2k/yr in costs. With the $28k system the un-adjusted payoff time is 6.7 years. A no-battery, no-UPS system is essentially maintenance free and should have a lifetime of about 30 years, so you get 23 years of "free" electricity afterward, worth (similarly un-adjusted for time value of money) about $96.6k.
Even when you correct for all the missing factors here, this seems like a pretty good deal.
(I live in Utah, where we have cheap electricity, no TOU charges yet, etc., so there's no economic payback to the solar PV I put in, or at least, not yet. Part of the reason I went for a small system, just 2.7 kW. So far, with roughly 10 months of actual operation, it's generated 3680 kWh. At our average cost of about 10 cents this is only "worth" $368, but at $.30/kWh it would be $1104. Scaled to 12/10ths those numbers are $442 and $1325, for yearly return. My end cost is a bit tricky, I had solar PV and solar hot-water installed together and the accountant worked the numbers as a group since they both get the 30% credit. But assuming $5/W, 2.7 kW = $13500 - $4050 tax_credit = $9450. It will take over 20 years of generation to pay it off here. But in Calif, with 3x higher rates, 20/3 is again about 6.7 years.)
(Our electric rates are going up though, we're already at .12/kWh in summer as of a few months ago. So now I get the summer generation, which produces the most energy, when the rates are highest, and the winter generation, which produces the least, at the lower rates. So my yearly payback estimate of $442 is too low. It's hard to be sure, but eyeballing the graph, I get maybe 60% of yearly production at the higher .12/kWh rate and 40% at the lower winter .09/kWh. Hm, let's see what the back of the envelope gets with those... about 4400 kWh/yr, 40% in winter = 1760 kWh @ $.09 = $158; 60% in summer = 2640 kWh @ $.12 = $317; so 1 year = $475, bringing the straight line number under 20 years. Interesting ... guess it all depends what happens to rates in the next few years, now.)