Suppose we wish to crack CO2 to C and O2. We need energy to do this. For simplicity, suppose the energy comes from burning carbon, since essentially our available marginal energy sources are of this nature. The laws of thermodynamics indicate that processes have less than 100% efficiency, so the amount of CO2 emitted from carrying out the cracking is larger than the amount of CO2 we crack, catalysed or not.

Essentially, given that we are burning carbon for our marginal energy needs, it would be better to use any energy we have to do something we need to do, rather than to crack the CO2. The only situation we would want to use energy to crack CO2 rather than do something useful is if it can be done using some energy that otherwise cannot be usefully used to do something. For example, some remote solar panels in the desert which cannot be practically joined to a grid might be set to crack carbon when they aren't doing anything else useful. But it might be more efficient for them to sequester it in some way instead.

This kind of argument also shows the absurdness of turning food crops into alcohol for fuelling cars. Much more energy can be got out of the land by growing biomass for powerstations. Since there are plenty of powerstations burning fossil fuel, we produce less carbon by fuelling cars with oil and burning the biomass in the powerstations. Of course if we can create cellulosic ethanol from cornstalks that's brilliant, but I suspect it would still be better just to burn them in a powerstation. On the other hand burning fossil fuels in cars produces localised urban pollution, which is lessened with ethanol, so on those grounds it might be better to have more cars burning ethanol.