"Energy can neither be created nor destroyed" is a nice motto. And has a certain ring to it. But it only tells you a misleading fraction of the total story. It turns out that some kilowatt hours of energy are worth considerably more than others. This is because of a very little known thermodynamic concept called exergy.
Exergy is a measure of energy quality. Exergy has a precise definition according to thermodynamics; with liquid fuels, exergy is related to a property called the Gibbs Free Energy. But real-world economics might give you a somewhat looser definition of exergy. Just by asking "how much is this stuff really worth?" For instance, electricity is just about the highest exergy stuff there is, as you can so conveniently move it or very efficiently convert it into other energy forms. Electricity often has a retail value near ten cents per kilowatt hour. Heat (especially at low temperature differentials) is much lower in exergy because of its gross inconvenience and its inefficiency in conversion to other forms. Because of this, those kilowatt hours in gasoline have a retail exergy value around three cents per kilowatt hour. Thus, a kilowatt hour of gasoline will usually be worth less than one third of a kilowatt hour of on-grid electricity.
Home electrical resistance heating gives a good example of the problems you get into with avoidable exergy drops. Using resistance heat (such as an electrical bar radiator), you get one low value heat kilowatt hour (kwh) of energy back for each high value electrical kilowatt hour input. Sell the electricity and buy natural gas, and you can get three kilowatt hours of heat energy back for each electrical kilowatt hour input. Better yet, run a heat engine backwards as a heat pump, and you can get five kwh of heat returned for each electrical kwh input. Chemical engineers go far out of their way to design processes that minimize loss of exergy. Any solar-to-fuel system which is to succeed absolutely, emphatically, and positively must avoid any large mid-process exergy drops. Because such drops can easily force any renewable and sustainable resource into becoming a net energy sink. Note that a process can appear to be fairly efficient and still lose so much exergy as to be useless. Electrolysis with its less than 1:1 conversion of high value kilowatt hours into low value kilowatt hours is an example.
According to IIASA, in 1990 the projected reduction of energy intensity by year 2020 was 36% in OECD countries, in reforming economies 42%, and in developing countries 11%. This means an average of 24% reduction in global emissions of CO2 per unit of economic activity during the 30-year period, from 290 grams of carbon per dollar to 220 grams. These improvements, while significant, would not be enough to offset the near-doubling in global economic activity over the period, estimated from US$24 trillion ($1012) to US$46 trillion (in constant dollars). Global emissions of carbon in the form of CO2 would rise from 5.5 gigatonnes of carbon in 1990 to an estimated 8.0 gigatonnes in 2020. Alternative estimates of carbon emissions in 2020 are 10.2 gigatonnes made by the USA EPA, and 9.8 gigatonnes used in the reference scenario of the Intergovernmental Panel on Climate Change (IPCC).