In the last video, we discussed the purpose of rebalancing, costs incurred during rebalancing, and we also chatted about a way to estimate those costs. But how do you know when to rebalance a portfolio? There are two types of events that trigger the rebalancing decision, cash flows and changes in the underlying parameters of the stock return model. Cash flows are movements of money into and out of a portfolio. For a portfolio of stocks, the most common forms of cash flows are dividends, capital gains and new contributions. For example, an investor receives dividend payments when he loans a stock. However, when he has shorted a stock, he is obligated to pay the balance of dividend payments issued to shareholders to the person who has loaned him the stock. Capital gains and losses are the increases or decreases in value of the assets in the portfolio, but these are unrealized until the assets are sold at which time they become realized capital gains and contribute to cash flows. New contributions consist of new investments in the fund or portfolio under management. The simplest way to adjust for cash flows is to buy and sell stocks using the current portfolio weights, but when some assets appreciate and some depreciate, the weights which represent the fraction of the portfolio invested in each asset change. A portfolio manager can take advantage of cash flows to adjust the under weighted or over weighted portfolios to return to the target asset allocation. Changes in model parameters do not automatically trigger rebalancing. In this case, the portfolio manager must decide whether the parameter change is large enough to warrant incurring transaction costs to rebalance. As a portfolio manager, you need to constantly monitor the parameters used in your portfolio and be ready to rebalance when needed. Parameters can change at anytime. For example, many types of corporate actions such as mergers and acquisitions or changes in management may impact the model’s parameters. When a portfolio manager suspects that a parameter has changed, she will re-estimate the model to come up with the optimal weights, then she will decide if the portfolio should be rebalanced by weighing the benefits of rebalancing with the downsides such as transaction costs. The most common rebalancing strategies are based on rebalancing at set temporal frequencies, or rebalancing when portfolio weights change by a certain threshold amount. One strategy is to rebalance at a set temporal frequency, daily, monthly, quarterly, annually or at another frequency, regardless of the changes in the sizes of the portfolio weights. This temporal frequency typically comes from the original time horizon of optimization. When a portfolio manager models his portfolio, he will often optimize it using data from a chunk of time of some duration. For example, if a portfolio is model based on month long chunks of data, the optimal weights are only valid for a month and so rebalancing is performed monthly. The threshold base strategy calls for rebalancing the portfolio only when the portfolio weights have drifted from the target ones by a predetermined threshold such as