With today’s CPI report continuing to paint a benign inflationary outlook, it is instructive to look at one component of that index, even if it is that component’s mismeasurments that create the true insights.
Plotted below are the index numbers and the year-over-year percent change in the Manheim Index versus the Used Vehicle-CPI. Clearly, there have been times when the Used Vehicle-CPI has diverged markedly from market reality. Note, for example, that the Used Vehicle-CPI was declining in the first four months of 2009. Talk about a false trend!
As can be seen in the graphs, the Used Vehicle-CPI will usually catch up to the trend in the Manheim Index after some lag. That is partly explained by the Used Vehicle-CPI not being a pure monthly series, but rather a 3-month moving average of depreciated prices. And those depreciated prices are based on guidebook data, which in turn has been “smoothed.”
The Used Vehicle-CPI, unlike the Manheim Index, is part of a cost of living index and, as such, must incorporate quality adjustments – a very difficult task. Nevertheless, one could argue that the Used Vehicle-CPI could be improved with better sampling, more robust data to preclude the 3-month rolling average, and a refined “expected depreciation” factor that now tends in instill a false cycle. For further information on the used vehicle CPI, I refer readers to Dr Pashigan’s paper: http://www.bls.gov/osmr/abstract/ec/ec010060.htm
And the BLS handbook:
But what are today’s mismeasurements telling us? If history is a guide, the Used Vehicle-CPI will now start to catch up to the Manheim Index. That will cause the overall CPI to be higher than it otherwise would be well into 2010. Additionally, since the base data in the Used Vehicle-CPI comes from the guidebook most used by lenders to set loan amounts, wide divergences from the Manheim Index quantifies one of today’s hot issues – i.e., lender loan amounts not keeping pace with inventory acquisition costs. Again, if history is guide, this problem should start to mitigate shortly.