"However, it should be noted that the above payback periods assume that the modules are always operated at their maximum power points [5], as with a maximum power point tracker."
And neglects the power reqired to constantly re-aim the cell(s). And apparently doesn't take into account degradation of cell efficiency over time (~1%/year), to name just a couple of criticisms via the referenced summary.
So maybe PV is close to having a net lifetime energy surplus. I do not doubt that it will someday (technology good). But for quite a while PV will be substantially more expensive than "mass produced" electricity, even accounting for distribution losses. Especially when considering small, distribued installations. Not amoritizing all of the other costs (brackets, tracking systems, inverters, grid interties and/or batteries) over a large customer base really drives the kWh cost up.