The arbitrage pricing theory and the intertemporal capital asset pricing model do not identify the risk factors that drive expected stock returns. Two approaches have emerged from the literature to fill the gap: macroeconomic models and fundamental models. A common view is that macroeconomic models have sounder economic rationale than fundamental models, but the latter fit the data better than the former. However, many papers in the U.S. have recently shown that even in empirical work, macroeconomic models perform better than fundamental models. This paper addresses this issue in the Tunisian stock market. We found no evidence that fundamental models explain average returns better than macroeconomic models in Tunisia. On the contrary, a macroeconomic model that contains as factors the term structure, private bank lending and retail sales generate the best empirical results. The empirical evidence here and in the U.S. implies that the common practice of using exclusively fundamental models in many areas of financial research is inappropriate. Macroeconomic models should be considered in the empirical asset pricing literature in addition to, or instead of, fundamental models.