is usually calculated by summing the market value of all physical noncash securities in the portfolio with the notional exposure of futures and forwards. If this notional value is greater than the market value of the total portfolio, the portfolio is notion-ally levered. For example, a portfolio with $99 million in physical holdings, Si million in cash, $5 million in a long S&P 500 futures position, and S3 million in a short U.S. 10-year bond futures is notionally leveraged as its notional exposure of $101 million is larger than its total value of $100 million.22 We prefer a measure of leverage that addresses whether the characteristics of the portfolio could have been obtained holding entirely physical securities. Under this definition, using derivatives to alter portfolio characteristics will not be flagged as leverage unless derivatives are used to create portfolio characteristics that would not be possible without the use of derivatives. For example, beta extension is easily obtained by holding stocks with betas higher than the benchmark, and duration extension is created by holding bonds with maturities longer than the benchmark, so strategies using futures to create a certain degree of these exposures would not be termed leveraged. In fact, such strategies are already employed by many traditional managers in their portfolios without being interpreted as leverage. THE EXPECTATIONS OF A GTAA PROGRAM Performance As noted at the beginning of this chapter, TAA and GTAA have experienced both strong and difficult performance periods over their histories. What should a client expect? The expectation for excess return is a function of the amount of active risk in GTAA as well as the manager's information ratio and investment style. Long-term information ratios on the best GTAA managers are generally between 0.5 and 1.0, and in some cases even exceed 1.0. While this may seem excessively high, consider that (1) transaction costs are 90 percent lower than for traditional products, and (2) unlike the tremendous number of players and amount of capital chasing after securities within a country, the number and capital of global asset allocators is relatively small. This implies that market disequilibriums are likely to be larger, and perhaps to exist for longer periods. This also suggests that GTAA managers may have longer periods of underperformance, although a well-diversified strategy should experience shorter episodes of underperformance. Performance comparisons across GTAA managers can be difficult due to the extremely customized nature of the strategy. However, for every GTAA account there exists a benchmark, so excess performance can be readily measured. We believe the information ratio on an AIMR-compliant asset-weighted GTAA composite is the best single performance measure, and this is readily compared across managers. We show the performance of our AIMR-compliant GTAA composite in Figure 25.4. The realized information ratio on this performance history is 1.14. 22The market value of futures contracts is zero because they are marked to market every day and the profit/loss is added to/taken from the cash in the account.