In 2006, Norway implemented a tax reform that implies that previously tax-exempt dividends now are taxed at a flat rate of 28 percent. Similar to portfolio investors, owners of closely held corporations can withdraw a tax-exempt amount of dividends (and c apital gains), the so-called Rate-of-Return Allowance (RRA). However, data show that many firms completely stopped paying dividends after the 2006 tax reform and pay less than the RRA as dividends (typically zero). Dividend payments recover only slowly an d still are way below the level prior to the announcement of the shareholder tax. As a result, the amount of unused Rate-of-return allowances that are carried forward has increased from around 5 billion NOK in 2006 to more than 11 billlions in 2007. We pr ovide a theoretical explanation for this behaviour, which we will test empirically by an examination of the financial policy of closely held firms before and after the implementation of the dividend tax with RRA. We will also test the financial lifecycle theory, using the ratio of earned vs. contributed equity and other proxy variables for maturity. Here, young firms are assumed to face ample investment opportunities but have limited financial resources. They will thus be less likely to pay dividends than mature firms, which have accumulated profits over some periods and will have abundant financial capacity. We will estimate the probability that firms do not pay dividends (and thus accumulate unused RRA that is carried forward), using proxy variables for maturity, and at the same time analyse the tax motives by looking at the relation between (no) dividend payments on one hand, and on the other hand the use of dividend substitutes such as repayment of original equity to the owners (or loans to the owner) . We use unique micro data for (changes in) individual owners' financial wealth in conjunction with accounts data of each closely held firm. Individual owners and their firms are matched by the shareholder register.