L_[CII]/L_IR ratio as a function of L_IR for local and high-redshift sources. Local (z < 1) sources are taken from Díaz-Santos et al. (2013). High-redshift sources are from De Looze et al. (2014) and Gullberg et al. (2015). Literature results for high-redshift quasars and companion sources are taken from Decarli et al. (2017) and Neeleman et al. (2019, for these we take an average of the reported L_IR). We note that we only consider the L_[CII]/L_IR as an upper limit for all companion sources. The error bar we used for BRI0952 comes from the uncertainty on the lensing factor on L_IR. The error bar on the L_IR and hence L_[CII]/L_IR ratio of AzTEC-3 is taken to be the range of L_IR values provided by Capak et al. (2011). For data points taken from other papers, we assumed L_FIR ∼ 0.75 × L_IR (following Decarli et al. 2017), but added an additional indicative error bar to the bottom left of the plot indicating the conservative estimate ∼30% due to this assumption (Carilli & Walter 2013).