Detection of Neptune-size planetary candidates with CoRoT data. Comparison with the planet occurrence rate derived from Kepler

[Abridged] Context. The CoRoT space mission has been searching for transiting planets since the end of December 2006. Aims. We aim to investigate the capability of CoRoT to detect small-size transiting planets in short-period orbits, and to compare the number of CoRoT planets with 2 \leq R_p \leq 4 Rearth with the occurrence rate of small-size planets provided by the distribution of Kepler planetary candidates (Howard et al. 2012). Methods. We performed a test that simulates transits of super-Earths and Neptunes in real CoRoT light curves and searches for them blindly by using the LAM transit detection pipeline. Results. The CoRoT detection rate of planets with radius between 2 and 4 Rearth and orbital period P \leq 20 days is 59% (31%) around stars brighter than r'=14.0 (15.5). By properly taking the CoRoT detection rate for Neptune-size planets and the transit probability into account, we found that according to the Kepler planet occurrence rate, CoRoT should have discovered 12 \pm 2 Neptunes orbiting G and K dwarfs with P \leq 17 days in six observational runs. This estimate must be compared with the validated Neptune CoRoT-24b and five CoRoT planetary candidates in the considered range of planetary radii. We thus found a disagreement with expectations from Kepler at 3 \sigma or 5 \sigma, assuming a blend fraction of 0% (six Neptunes) and 100% (one Neptune) for these candidates. Conclusions. This underabundance of CoRoT Neptunes with respect to Kepler may be due to several reasons. Regardless of the origin of the disagreement, which needs to be investigated in more detail, the noticeable deficiency of CoRoT Neptunes at short orbital periods seems to indirectly support the general trend found in Kepler data, i.e. that the frequency of small-size planets increases with increasing orbital periods and decreasing planet radii