### Table 7 Hedging Demands with Learning Panel A: Hedging Demand

2005

"... In PAGE 35: ... The revisions in the beliefs about the return generating process, which determines the future investment opportunities, are strongly correlated with returns, inducing additional hedging demands. Table7 quanti es these e ects. Panel A shows the di erences between the dynamic and myopic allocations.... In PAGE 36: ... This result again illustrates the importance of incorporating learning into the dynamic portfolio choice. The most striking result in Table7 is that, in the case of parameter uncertainty with learning, the hedging demands are negative and even further decreasing with the horizon. This pattern in the optimal allocation is attributable to learning about model parameters.... ..."

### Table 5: Hedging with options in a MJD framework

2004

"... In PAGE 14: ... As in Section 3 we also investigate the sensitivity to interest rate risk in a model with jump risk, which is a more appropriate framework for the price dynamics of the underlying of a lookback option embedded in a GMDB with ratchet feature. Table5 illustrates the performance of hedging with options in a MJD framework. Table 5: Hedging with options in a MJD framework... ..."

### Table 5 Comparisons between hedging models

"... In PAGE 15: ...uch an approach is 0.956218, very close to unity. We should mention here that although the ECM with GARCH (1,0) errors performs statistically better than the simple error correction representation, it did not manage to increase hedging effectiveness, as measured by using the adjusted R2, and hence the simple error correction representation is considered superior. Table5 summarizes the comparisons of the optimal hedge ratios estimated using alternative methods. In terms of risk reduction, the appropriate method for estimating optimal hedge ratios is the ECM.... In PAGE 15: ... 5.4 In-sample analysis Table5 reports also the RMSEs, MAEs and MAPEs for each model. The results indicate that the error-correction specification outperforms all the other models since ... ..."

### Table 4 - Hedging test results

2002

"... In PAGE 22: ... Hedge ratios are calculated using the standard technique. Table4 reports the hedging test results. The results of using the three methods to hedge a written $100 OTM put position using only the underlying asset are quite interesting.... In PAGE 23: ... When the t-statistic is significant and positive, the daily EPK hedge performance is significantly better than the alternative model hedge performance. The fourth column of Table4 reports this measure. The t-statistic for the daily EPK model outperformance is 3.... In PAGE 32: ... Table4 reports the hedging test results for the 912 sample dates over the period 1991 - 1995. The particular hedging problem chosen is hedging a $100 position in out-of-the-money (OTM) S amp;P500 index put options using at-the-money (ATM) put options, the S amp;P500 index portfolio, or both.... ..."

Cited by 9

### Table 10. Results of Hedging Experiment

"... In PAGE 24: ... The hedged net cash flow is the sum of the unhedged net cash flow and the profit or loss on the hedging activity. Table10 summarizes the results of the hedging experiment. For each model, several statistics are computed for each exchange rate.... ..."

### Table 5: Hedging `Violations apos;

"... In PAGE 34: ... (1999). In Table5 , we compare the frequency of observed Type I and IV violations in S amp;P 500 options in the period between 1986:4 and 1996:5, and the frequency predicted by the U-Model, using the deltas shown in Corollary 7, the belief process shown in Figure 7 and the moneyness of each traded option. As seen for the full sample, the model apos;s predicted level of Type I errors for both calls and puts are quite close to the empirical proportions, but its predictions of Type IV errors fall short.... ..."

### Table 2: Benchmark results edge crossings h-edge crossings time/s

"... In PAGE 13: ...07 Z9sym 3901 2320 1788 16 summationtext 11933 8273 6179 32.91 In Table2 it is shown that fewer edge crossings in the graph embedding lead in almost all cases to less crossings in the orthogonal drawing if only the sifting algorithm is applied. To obtain less crossings in the graph model we postprocessed the results computed with the averaging heuristic method with the windows optimization procedure [15].... In PAGE 14: ... The runtime for the sifting combined with the reordering algorithm are published in the last column. In the last rows of Table 1 and Table2 we give the total number of... ..."

### Table 3. Comparing the accuracy and number of high information instances found in a 3 category decision tree model with quality uncertainty.

"... In PAGE 11: ... Here, we divided the training data into a High category where information helped 10 or more units, a Low category where information helped less than 10 units, and a Zero category where information did not help. The results of this experiments from the quality set are found in Table3 , and those from the duration experiments are in Table 4. Within these tables, we present the classification confusion matrix, often pre- sented in multi-category classification problems.... ..."

### Table 4. Comparing the accuracy and number of high information instances found in a 3 category decision tree model with duration uncertainty.

### Table 3. Hedging effectiveness under risk minimization

"... In PAGE 6: ... We note, for example, that the mean futures return is significantly positive under the ECT model in Table 2. Table3 presents in- and out-of-sample comparisons of the hedging effectiveness of the CONST, ECT, and GARCH-ECT models under the risk-minimization criterion implied by the first term of equation (4). The table contains the mean portfolio return and portfolio variance for hedge ratios that minimize portfolio variance.... ..."