|
The REWARD_CLASS_PRICEMOVEMENT_YIELDREVERSION is comprised of the following elements:
Each of these elements seeks to measure the degree by which the issue's yield (as measured in the manner appropriate to the element) will revert to its mean historical value. For further information on what is meant by the term "prior values", see the glossary items instrumentAveragesRecord and exponentialMovingAverage.
The following graph shows the relationship between the value of REWARD_CLASS_PRICEMOVEMENT_YIELDREVERSION and the ultimate valuation that HIMIPref™ assigns to each instrument:
| Regression Analysis |
| Intercept |
4.1266 ± 4.0612 |
| Slope |
-0.9421 ± 2.3682 |
| Regression Sum of Squares |
108.6062 |
| Residual Sum of Squares |
686.2147 |
| R Squared |
0.1366 |
| F Statistic |
6.0933 |
| Rejection Threshold (SDev) |
3.00 |
| Rejection Count |
6 |
| Observation Count |
79 |
For example, when evaluating PWF.PR.E on 2005-12-30:
An interesting feature of the current (2006-01-26) parameterization of HIMIPref™ is that the larger and more meaningful numbers (instrumentPortYieldReversionSpeed and instrumentCostYieldReversionSpeed are positive (which implies that momentum is operative), but the slope of the correlation is negative. This "double negative" implies that reversion to mean is, in fact, the operative mechanism but that this group of parameters acts as a correction factor to other, more important, reversions.
The following graph shows some support for this view: there is a significant negative correlation between REWARD_CLASS_PRICEMOVEMENT_YIELDREVERSION and REWARD_COMPONENT_REVERSION_PRICEDISPARITY, which is strongly correlated with total valuation (see Reward Class Disparity).
| Regression Analysis |
| Intercept |
0.8104 ± 1.1844 |
| Slope |
-0.7078 ± 0.6906 |
| Regression Sum of Squares |
115.9461 |
| Residual Sum of Squares |
86.0176 |
| R Squared |
0.5741 |
| F Statistic |
53.9174 |
| Rejection Threshold (SDev) |
3.00 |
| Rejection Count |
7 |
| Observation Count |
82 |
|
