Despite the intense effort developed over the past 10 years to determine the ¹²C / ¹³C isotope fractionation associated with photorespiration, much uncertainty remains about the amplitude, and even the sign, of the ¹²C / ¹³C isotope fractionation of glycine decarboxylase, the enzyme that produces CO₂ during the photorespiratory cycle. In fact, leaf gas-exchange data have repeatedly indicated that CO₂ evolved by photorespiration is depleted in ¹³C compared with the source material, while glycine decarboxylase has mostly favoured ¹³C in vitro. Here I give theoretical insights on the glycine decarboxylase reaction and show that (i), both photorespiration and glycine decarboxylation must favour the same carbon isotope — the in vitro measurements being probably adulterated by the high sensitivity of the enzyme to assay conditions and the possible reversibility of the reaction in these conditions, and (ii), simplified quantum chemistry considerations as well as comparisons with other pyridoxal 5′-phosphate-dependent decarboxylases indicate that the carbon isotope fractionation favour the ¹²C isotope by ~20‰, a value that is consistent with the value of the photorespiratory fractionation (f) obtained by gas-exchange experiments.