The transit method is a powerful means of detecting and characterising exoplanets. The Kepler spacecraft monitored 150,000 stars over 4 years, producing light curves for more than 4000 planetary candidate systems. Many of these appear to be multi-planet systems, consisting of between 2 and 7 planets. The relative frequency with which systems of differing multiplicity are detected depends on the underlying multiplicity distribution of exoplanet systems, and on the relative inclinations between the planets that make up the multiple systems. It is of interest to determine the underlying distribution of mutual inclinations, as it provides information about the formation and evolution of the planetary systems.

Following Johansen et al (2012), we have developed a method for constructing synthetic multiple planet systems that uses the distribution of semi-major axes from the Kepler data, and which allows the mutual inclinations to be varied systematically. By generating and “observing” synthetic planetary systems, we have undertaken a statistical analysis to determine which values of the mutual inclination give rise to a synthetically detected multiplicity distribution that agrees with the Kepler data. Reasonable agreement with the Kepler data was found for mutual inclinations between 3 and 8 degrees, indicating that the inclinations within the Kepler systems are small, consistent planet formation within a protoplanetary disc. In agreement with earlier work, however, we find that there are too many single planet systems within the Kepler data to be consistent with the idea that all Kepler systems are multiple systems with a simple distribution of mutual inclinations.