Case Study: Shell International

In March, FAS133.com spoke with Nick Grantley, FAS 133 treasury project manager at Shell International to get his practitioner’s view of the larger FAS 133 issues and a practical implementation approach. As the Q&A illustrates, Shell adopted a rather straightforward approach to testing effectiveness relying on cumulative dollar offset for both prospective and actual testing.

While, as Mr. Grantley notes, this approach avoids many of the pitfalls (in terms of validity) of statistical testing, it does have some drawbacks. In particular, the dollar-offset approach may well “disqualify” more hedges since it’s mathematically more sensitive to price changes. Still, as Mr. Grantley says, “consideration on which effectiveness test to select comes into its own when the hedge itself is ‘borderline.’” Shell’s hedges, he contends, do not fall into that category.

FAS133.com:  Is FAS 133 specific enough, or overly specific, in its requirements for effectiveness testing?

NG: In order to perform effectiveness testing for Shell Treasury, the standard’s guidelines and the clarifications issues by the DIG are sufficient.  The highly effective boundary is clear and the requirement to test effectiveness at least quarterly and when financial figures are reported (which, for Shell, is quarterly) are clear. 

FAS133.com: There is much discussion in the market regarding the 80-125% rule and other statistical boundaries, but there is not necessarily a boundary mentioned in FAS 133. How do you define a highly effective hedge?

NG: From the Shell perspective we are clear: We identify the risk that we want to hedge.  If the identified risk changes, then the value/cash flows of the hedged item should be offset by approximately an equal and opposite move in the hedging item. The tolerance boundary for this comparative move, past and projected is 80 – 125% (in order for the hedge to be deemed highly effective, and therefore, qualify for hedge accounting).  I have not explored the other boundaries that you allude to, as the 80 – 125% is acceptable for our measurement purposes. 

FAS133.com: What are the available methods, and what have you chosen as your approach for calculating prospective effectiveness?

GN: The Shell approach is to keep what is a complex issue as simple as possible. For our fair value hedges we have elected cumulative dollar offset (for both actual and prospective testing), which has the edge on dollar offset as it smoothes any “ineffectiveness”. Our approach is to create theoretical Net Present Values (for the hedged items and the hedges) using the most current yield curve. This method is one of the few that is forward looking (by using market forward/future rates); and yet, by comparing past results has the 'retro' angle. The advantage is that this approach is not statistically driven (statistical results are generally not widely understood) and it avoids judgmental issues on confidence intervals, data ranges etc.

FAS133.com: It sounds like you use cumulative dollar offset for retrospective analysis and actual measurement of effectiveness, but then a theoretical NPV (for both hedge and underlying) using available yield curves to determine prospective effectiveness. Do you combine the two at all? Iin looking forward, do you also take into account the past performance? And, if so, how do the NPVs and dollar offset figures mesh?

NG: Yes. This is where the distinction between retrospective and prospective effectiveness converges. Our retrospective, dollar offset looks at changes in fair value/cash flows arising from past events. Our prospective testing looks at changes in fair values/cash flows in the future. In addition, we have the prospective considerations documented at inception and transitions that are reviewed regularly to ensure they still hold true.  This is narrative (text) on the future hedge logic demonstrating ongoing hedge effectiveness.

FAS133.com: What has been your experience of the sort of methods offered by software vendors?

GN: Our software house, SunGard, provides regression, dollar offset or shortcut method as the three alternatives.  The feedback that I have received is that, in practice, the criteria for the shortcut method is hard to satisfy—particularly the requirement to show that the NPV at inception is zero and also the problem if the variable leg of the interest rate swap is different from the benchmark interest rate being hedged. The impression I have is that many providers offer a similar test solution. An accountancy consultant also told me that they haven't yet seen a system that correctly handles the release of OCI.

FAS133.com: Is it true that the only widely available statistical methodology is regression? Are there any alternatives, and do you see drawbacks to regression?

NG: Although regression is one of the less sexy options (as everyone has heard of it!), it is a pragmatic solution to testing effectiveness and in a way demonstrating correlation it what it’s all about.  I think it is a good method. Consideration on which effectiveness test to select comes into its own when the hedge itself is “borderline.” I can’t think of any examples where a perfect economic hedge would give an ineffective measurement result, outside 80 – 125%—I’m assuming an appropriate choice of designation and benchmark.  The Shell Treasury approach is to put in place highly effective economic hedges, and therefore, we have selected simple methods for demonstrating this.

FAS133.com: Is it possible to approach FAS 133 effectiveness testing from both an accounting and risk management standpoint?

NG: When I was first undertaking the effectiveness hedging, I found a difference between fair-value changes in a swap and a borrowing. Investigation showed it was the conflict between hedge effectiveness and fair value accounting.  The fair value of an instrument includes the interest accrual.  In one example involving a bond and an interest-rate swap, the bond had one accrual in the valuation, while the swap had two—the additional accrual can materially affect the result and is where accounting and risk management diverge.

The approach I adopted was to exclude the accrual from hedge effectiveness, although include it for the main fair-value accounting. I suspect, in practice, most economically (whn is likely to be more academic!  

FAS133.com: Is the exclusion of accruals in line with FAS 133’s rules?

GN: To tell you the truth, I don’t know.  From a pragmatic point of view, very few interest-rate hedges would remain in the 80 – 125% boundary if the interest accrual was included; and therefore, I excluded on that basis.  Examples I have seen ignore the interest accrual, but the authors seem to exclude interest accruals more for simplicity rather than a definitive ‘this is how to do it.’  I would be interested to know what the readers think.

FAS133.com: Do you foresee increased income statement volatility (in general if you feel uncomfortable making perditions about your organization) as a result of this? 

NG:  Across many companies I suspect the answer is yes.  By marking to market derivatives and revaluing underlying exposures in relation to the risk being hedged, the two are unlikely to equal.  In addition, OCI volatility needs close scrutiny.  Within Shell, FAS 133 will not have a significant effect on the group’s reported net earnings. 

There will be some treasury structures that will appear less attractive to undertake because of their income volatility effect (e.g., cross currency interest rate swaps into non-functional currency).

FAS133.com:  Do you think some multinationals will simply mark to market their hedges to avoid the hassle of testing and making appropriate accounting entries? In other words, is the OCI "halfway house" worth all the hassle? And, when does hedge accounting really make a difference?

NG: Where there is a limited impact I suspect hedge accounting would not be worthwhile. The use of OCI in cash flow type hedges is an added complication but one that is currently over-estimated relative to fair value hedges.  The long-term solution must be to revalue all financial instruments. When systems have aligned themselves to do this, it will be far less burdensome for fair value hedging.