Drafting a 2016 year end summary report format for PBC. The numbers below are just old NIM levels for 2015 and formatting estimates. PBC will show the actual estimates from the Jan 2017 run.
The 2016 year end PBC batch will compute the aggregate optimal NIM and allocate back to the top Bank Holding Companies both the optimized annual NIM and the estimated end of year credit exposure. Then we will breakdown the actual NIM explanatories to Credit, Interest Rate, Execution, and Unexplained (all in actual NIM basis points). I think one interesting thing about these numbers is to fit the Capital plan implementation Execution model back to the actual realized NIM.
Let’s go over how these numbers can be generated. The aggregate Actual NIM for 2016 will presumably be available for the US BHCs in Jan 2017. The US Federal Reserve will disclose the monthly/quarterly composition of Assets and Liabilities held in aggregate across all the US BHCs on Jan 2016 and Jan 2017. PBC will run the multiperiod LP on these aggregated securities to determine the Optimal NIM for 2016 and the Optimal Credit Risk on the aggregate portfolio on Jan 2017. The optimal Credit Risk is an interesting number because it represents the risk left on the books needed to achieve the optimal NIM (note: we probably want to back out an estimate to the realized Credit01 on the books on Jan 2017 for comparison with the optimized CR01). PBC then runs another LP to fit the period by period NIM of each of the named BHCs and other. We know the NIM and the historical product cash flow runoff so we compute the optimal assignment of products to BHC that reproduces the observed (reported) NIM behavior. Then we run the Optimal NIM and Optimal Credit01 to fit the aggregate numbers. That gets you everything upto the Attribution of the Actuals on Jan 2017.
Let’s use the aggregate portfolio configuration to compute the Credit, Interest Rate, and Reinvestment 2016, attributions for Aggregate line item. The Reinvestment 2016 allows PBC to impute a Capital Allocation plan (in aggregate) and then run that plan though the LP to find the optimal schedule for that plan. Use that number to estimate the Aggregate execution cost. The aggregate Unexplained will be the difference between the Realized NIM and the sum of the Attributed NIM. Finally on an individual BHC basis we rerun this process to get the Credit, Interest Rate, Reinvestment, and Execution Attributions that best fit the now known aggregate parameters. We back out the individual BHC unexplained components from the difference with the published Actual NIM 2016.