An incandescent bulb produces about 5% light and 95% heat. CFLs are said to produce 25% light and 75% heat. It is this little difference which constitutes the foundation for all other figures and calculations.
However, as shown earlier, top brand CFLs only give about 3-4 times more light at best, while many CFLs on the market give less light, and proportionally more heat.
Update Aug 2012: updated table.
In cooler climates such as in North Europe or Canada, where it is usually cold and dark at the same time, the “excess” heat from lamps naturally adds to indoor heat and is thereby not necessarily wasted, even if not all of the extra heat affects thermostats enough to lower heating bills or is produced during the heating season.
• A (pro-CFL) study by the Swedish Energy Agency 1998 showed varying results depending on type of house, heating system, thermostat efficiency, season, latitude, amount of direct sunlight etc.  “Very roughly [with seasonal variations included] one may expect a net saving of 50% of the lighting savings in a house heated by electricity.” 
• According to a British 2003 study, about 60% of the energy from lighting throughout the year in a typical British house turns into useful heat  (as 60% of the energy lightbulbs consume heats the building they are in so heating needs increase accordingly) with CFLs saving only 20%. In a follow-up study using thermal simulation software, researchers conclude:
“The findings from this study confirm the earlier conclusions that the HRE is a significant factor and therefore one that needs to be taken into account to obtain realistic predictions of the savings from reducing energy consumption by lights and appliances within buildings.
“In a typical UK house, the cost saving from installing low energy lighting, if the HRE is ignored, will be overestimated by about 19% and the carbon saving by about 67%. It would be reasonable to expect a similar level of overestimation when looking at the potential savings for a large group of dwellings, rather than an individual typical house. Failure to recognise this when performing calculations could lead to wrong conclusions being drawn and, potentially, to wrong decisions being made. [emphases added] 
• Recent Canadian studies suggests that actual savings of potential savings depends on season and what type of energy you use for electricity. In some regions it may even be counter-prodcuctive. [5, 6, 7]
“Physics department head Peter Blunden found using CFLs in Winnipeg could cut energy consumption by 67 per cent, “but that’s not the whole story,” he said. “The issue is all the heat that’s thrown off by the incandescents.” Blunden said factoring in heating and cooling changes, Winnipeggers would end up with energy and cash savings of 17 per cent, similar to Manitoba Hydro’s findings. Those who use air conditioners would see savings of around 24 per cent, he said, while cash savings will be a little higher for people who heat with gas instead of electricity.”
Blunden pointed out that lights make up a tiny portion of a home’s energy needs, just three per cent on average. ‘We’re really talking about a very small slice of the energy pie,’ he said.” [emphases added] 
• Dr Peter Thornes explains the heating benefit in more detail here: A Heat Benefit and makes the astute observation that it’s funny how the excess heat from lighting seems to be considered a highly relevant factor when it comes to space cooling when it’s too hot, but not when it comes to space heating when it’s too cold, although both are two sides of the same coin! 
Faced with these annoying facts, whenever the heat replacement effect is mentioned, CFL proponents have been trained to automatically retort that “incandescent lamps are inefficient to heat houses with.” This argument rather falls on its own ridiculousness as I’m sure no one would dream of turning on a light bulb just to create heat! Lamps are obviously used for lighting houses with, and the extra heat is just an added bonus.
• Another early argument was that the light bulbs “have to be placed under windows” in order to have the same heating effect as radiators in reducing cold draft. But why on earth would one place a light bulb under a window? Drafty houses usually have a radiator there… No one has claimed light bulbs can replace radiators, only complement and decrease the need of some of the heat they give off (a very tiny portion at that).
• With increasing desperation, critics now claim that the heat “only stays near the ceiling” but this is not true either as heat circulates, and most people use floor-, table and desk luminaires besides ceiling- and wall fixtures.
Getting a little extra heat close to where one is sitting is usually an immediate benefit whether themostats register it or not – except during hottest season when it may be a nuisance instead (not a huge problem here in Sweden where the slightly-too-hot season usually lasts about 2-6 weeks and coincides with the little-need-for-light-anyway-since-the-friggin’-sun-never-sets season).
The warmer the climate you live in, the less of a benefit and the more of a problem incandescent heat will be, of course, and it so happens that warm incandescent light is most popular in the cold and dark climate zones and decreasingly popular the closer to the equator you get. Perhaps there is a natural reason for this? Seems like consumers are alredy intuitively drawn to the type of light that is most appropriate for their particular climate, so why regulate with a one-size-fits-all solution that will be an ill fit for many?
1. Studie över spillvärme från hembelysning, Enheten för Energiteknik, 1998
2. Beräkning av energibesparing vid byte till lågenergilampor, Energimyndigheten (STEM), 1998 (Swedish study)
3. Market Transformation Programme: The Heat Replacement Effect (UK study)
4. Thermal modelling of the heat replacement effect and its implication for energy saving programmes (UK study)
5. Benchmarking of energy savings associated with energy efficient lighting in houses.pdf (Canadian study)
6. “Switching off incandescents a no-brainer?”
7. “Compact bulbs not as green as once thought”
8. New Electric Politics: A Heat Benefit
• Lighting only uses around 3% of a household’s total energy consumption.
• Used indoors in cooler climates during the heating season, only a part of nominal savings from switching to CFLs will be real savings, due to increased need for space heating to make up for the heat no longer produced by light bulbs.
• How much depends on many factors such as length of heating season in one’s particular climate zone, house type, insulation, type of energy used in ones region or utility, energy system (e.g. gas, heat pump, electric radiators, water radiators), presence of indoor thermostats and thermostat sensitivity – though nothing stops the home owner from turning down the heat manually if not sensitive enough, and enjoying both the warm light and little extra heat from incandescent and halogen lamps instead (if one needs the light anyway).
• Only in warm climate zones and seasons which require extra cooling do the most efficient CFLs potentially save what they are claimed to save.