Covering 20 years (1995–2015), the article ascertains the presence of the month-of-the-year effect in the Indian stock market, for the raw returns series as well as after adjusting for non-linearities of the market. Whether the effect is the same for portfolios of different sizes and values is also ascertained. The threshold generalised autoregressive conditionally heteroskedastic (TGARCH) model is employed to address non-linearity. The results suggest the presence of higher returns in November/December at the index level. Further, only firms with a size smaller than the average exhibit seasonality in the form of the April/May and November/December effect. The value-sorted portfolios exhibit weaker evidence of the December effect. Tax-loss selling, window dressing and behavioural aspects seem to provide the explanation.