MI

Management Information

The trend tracker is used to identify outlier behaviour that appears within each individual time series. All time series across all time periods are compared to their historical movements to identify anything out of the ordinary. This both alerts MIAC to any issues within the model outputs for any given month and therefore serves as a useful MI tool for model validation and governance and provides an automated way to identify genuine trends that exist within the data.

Here we detail some of the highlights of the MI analysis that are performed on a monthly basis, their rationale along with possible actions clients could perform to mitigate any flagged weaknesses.

Index Stats.

Price Stability

We closely monitor the cross-temporal stability of the MIAC HPI to ensure outliers are flagged for clients so that they may be able to explain collateral price volatilities within their portfolios. This would be a requirement for any index particularly one that provides such granularity. Clients may decide that they are happy with these tail events (as either they are not considered too extreme or they concern a very small proportion of their portfolio) or that they wish to dampen this volatility via some smoothing methodology. In order to discern whether or not the previous months' change in average price is an outlier we calculate the percentile to which it belongs within the historical distribution of price changes for that given County, London Borough, Local Authority/Property Type. If the distribution of price changes suggest that the most recent price change is a tail event with a probability less than or equal to 5% then we flag this as an outlier and investigate accordingly.

Percentile Outliers

Regional Spread

One of the advantages of using the MIAC HPI over other indices is the extra granularity of house prices provided by the geographic level of either county, London Borough, or Local Authority. In order to illustrate the benefit of this extra granularity we chart the distribution of County, London Borough, and Local Authority indices within each region and provide the MIAC Regional price for comparison.

Missing Values

Missing values are where there are too few data points in a given County, London Borough, Local Authority/Property Type on which to base an estimation for a given month. In which case the model reverts to the less granular regional level in order to produce a price. Whilst the fact that the number of data points within certain segmentations are low means that the same is likely to be true within a given portfolio, it is however instructive to track the time series for which 'missing data points' exist so that clients are able to understand and monitor these restrictions internally. The time series with missing data points for series with no data points are known over the time series Time series that fall into the first category are listed within the analysis and for those that fall into the second we chart the number of consecutive missing data points from the current month.

Further Analysis

Further Analysis If there is additional analysis that a client requires please reach out and we can attempt to include this within future MI.

Request a Demo
Access R-AVM >>