We propose a novel explanation for classic international macro puzzles regarding capital flows and portfolio investment, which builds on modern macro-finance models of experience-based belief formation. Individual experiences of past macroeconomic outcomes have been shown to exert a long-lasting influence on beliefs about future realizations, and to explain domestic stock-market investment. We argue that experience effects can explain the tendency of investors to hold an over proportional fraction of their equity wealth in domestic stocks (home bias), to invest in domestic equity markets in periods of domestic crises (retrenchment), and to withdraw capital from foreign equity markets in periods of foreign crises (fickleness).Experience-based learning generates additional implications regarding the strength of these puzzles in times of higher or lower economic activity and depending on the demographic composition of market participants. We test and confirm these predictions in the data. * We thank Vladimir Asriyan, Fernando Broner, Nicolas Courdacier, Jaume Ventura, and seminar participants at CREi-UPF and the NBER ISOM conference 2019 in London for their comments and suggestions, and Clint Hamilton for excellent research assistance.future stock market returns, and this long-term influence is more pronounced among younger generations (Malmendier and Nagel, 2011). 2 Models of experience effects naturally generate both the overweighting of recent experiences and the long-lasting effects of lifetime experiences. Experience-based learners assign extra weight to realizations of macro-financial variables that they have personally experienced when they form beliefs about future outcomes of the same variables. A given crisis experience exerts stronger influence on younger cohorts, for whom the crisis experience constitutes a larger portion of their lifetime histories so far. 3 In addition, models of experience-based learning also rationalize classical asset-pricing puzzles such as return predictability (Campbell and Shiller, 1988, Fama and French, 1988) and excess volatility (LeRoy, 2005, LeRoy and Porter, 1981, Shiller, 1981, and micro-level stylized facts such as investors chasing past performances. 4These insights have direct implications for international macro models, as investors in different countries have different "experiences." They differ in their exposures to domestic versus foreign outcomes, and different countries also have different demographics. For example, applied to the Asian financial crisis, experience-based learning would predict that the crisis has exerted a significant influence on the risk attitudes and beliefs of East Asians, and its influence was different for, say, Europeans -even controlling for income, wealth, and other standard factors. Moreover, comparing regions across and within East Asia, the influence is predicted to differ depending on where a higher fraction of young market participants were exposed in each country. Nevertheless, most of the international finance models to date ar...