Questionable Statistical Methods Used for Important Policy Finding

By Jan Galkowski posted 10-19-2015 22:13

  

The disposition of requests for building additional methane capacity for heating and electricity generation has fallen within the pursue of the Massachusetts Department of Public Utilities ("DPU"). This involves supporting and directing contracts between public utilities, such as National Grid and Eversource, and fossil fuel providers, such as Spectra/Algonquin, which have the net effect of authorizing the build of additional fossil fuel infrastructure, such as pipelines and electrical generation plants, in Massachusetts, with 30-50 year depreciation schedules, and all these being paid for by passing the costs on to ratepayers.  The order authorizing such is available.

 Notable in this order is the recapitulation of the DOER argument, that "DOER points to the 99 percent positive correlation between regional average monthly gas and electricity prices in the winter months to support its view that electricity prices in the winter of 2013-2014 reached unprecedented high levels as a result of the gas price spikes (DOER Comments at 4, 22; DOER Reply Comments at 3, 15)", along with a DPU footnote explaining what a correlation coefficient is. (It does not explain what a correlation coefficient is not.) This is taken by both DOER and DPU as a conclusive attribution of the increase in prices to the gas price spikes and, further, in the sequel, to the shortfall in methane capacity.  Other commentators to the discussion, including the ISO New England entity, responsible for electrical stability and supply to the New England grid, interprets matters differently

The Energy Information Administration has long observed that methane prices in New England were higher than in other markets, but has shown a more nuanced interpretation of the present situation. In particular, its interpretation does not assign responsibility to pipeline capacity exclusively for elevated methane and, hence, electricity prices.

 Concurrently, the Union of Concerned Scientists ("UCS") has issued a report The Natural Gas Gamble, which seriously questions the propensity of various states to pursue a short term solution via methane for greenhouse gas targets and energy shortfalls. Both UCS and other participants in the discussion, such as the Conservation Law Foundation ("CLF") indicate there is ample historical evidence that the connection between supply and prices is more convoluted, and that the long term risks of permanently acquiring large fixed assets predicated on a hasty determination of this connection, especially in terms of energy prices, warrant a more careful approach. The Attorney General of the Commonwealth of Massachusetts agrees.

 Accordingly, I submit for your collective consideration, the premise that significant societal determinations ought to be made using the best statistical practice possible, that the DOER/DPU methodology is a grossly flawed statistical practice, that it does not formally incorporate all the perspectives and knowledge available from the contributors to the discussion, that quantitative means of doing so are, in fact, available to the student versed in the art, and that, therefore, in accordance with the Ethical Guidelines for Statistical Practice of this organization, suitably qualified statisticians with an interest in this area and jurisdiction, ought to indicate the shortfalls of the analysis in question, and guide the discussants towards a remediation.

 Thanks.

  -- Jan Galkowski, quantitative engineer and statistician, Westwood, MA.

0 comments
230 views

Permalink