Do the individual preference indicators developed in the previous papers ­ scores, scales, isolated questions, lottery choices - have any significant effects on net or gross financial assets in keeping with the predictions of microeconomic savings models? The empirical analysis of the 1998 INSEE Patrimoine survey on personal wealth comes to a positive answer to this question, but especially in the case of the scores: obtained from a multiform questionnaire, these composite indicators have much greater and more consistent explanatory power than that of the other preference measures considered. Being more prudent or more provident increases the sum of wealth held. Family-related altruism also goes hand in hand with greater wealth. Households' assets therefore definitely appear in all their forms: precautionary reserves, lifecycle savings, especially for old age, and transmission to family members. However, the extent of impatience, a composite indicator of short-run reactions, has no effect on the level of accumulation. The same holds true for non-family related altruism (concern for the welfare of future generations in general, for the future of the planet, etc.). The explanatory gain obtained with the preference scores may seem modest due to the extremely high concentration of wealth. Unobserved heterogeneity is not greatly reduced. Yet despite all this, the quantitative effects of these subjective variables are far from negligible, especially as regards the time preference: between individuals at either end of the scale ­ i.e. between the shortest sighted and the farthest sighted in the sample ­ the estimated deviations in personal wealth range from 1 to 10.