Interpretation of the stress testing results

  • 759 Views
  • Last post 28 November 2019
daringodber posted this 20 November 2019

Hi, I am trying to interpret the Stress testing – summary report. If I look at the cashflow measure for the capacity test for Q4 2019 it tells me the weighted average is .027, the low is -.090 and the high is .060. My interpretation from this is that the weighted average impact of cash flow impact is 2.7%. This 2.7% is the change in cashflow divided by the cashflow from last set of accounts. So as an example, if the weighted average cashflow form all participants accounts was $100 then the cashflow impact would be +$2.70. Using this same example, the low (which I assume is the lower quartile) would be -$9.00 and the high (which I assume is the upper quartile) would be +$6.00. Am I interpreting this correctly?

Order by: Standard | Newest | Votes
Callum M posted this 28 November 2019

Hi Darin

My apologies for the time to respond to your query. Your question has arisen because the data tab for the 'stress testing - summary report' was poorly labelled. The team has made several improvements which should help clarify what's what. At the heart of the misunderstanding is that the EMI data tab used to have a column labelled 'weighted average'. That has since been relabelled as 'sector aggregate'. There's a good explanation of the sector aggregate in the file attached to this post - https://www.emi.ea.govt.nz/Forum/thread/stress-test-reports-now-on-emi/ - but I'll quote the relevant bit here:

"A sector aggregate is computed for each stress measure and is indicated on the EMI reports with a black diamond. The sector aggregates are computed by summing the numerators of all disclosing participants and dividing this by the sum of the denominators of all disclosing participants. The numerators and denominators are those used in the formulae above in this table. Note that these aggregates take no account of the size of any one participant relative to another and should not be interpreted as averages."

As you'll see from that explanation, the sector aggregate is not a weighted average. When compared against the median result in each quarter, the sector aggregate does provide an indication of the size of the parties above or below the median result. In nearly all quarters, the sector aggregate is above the median.

Regards,

Callum McLean

 

daringodber posted this 28 November 2019

Thanks Callum. That didn't answer my underlying question about the interpretation though.

If I look at the cashflow measure for the capacity test for Q4 2019 it tells me the lower quartile is -.050. My interpretation from this is that the lower quartile cash flow impact is -5%. This 5% is the decrease in cashflow divided by the cashflow from last set of accounts. So as an example, if a participants accounts was $100 then the cashflow impact would be $-5. Am I interpreting this correctly?

Also, between my original post and this post the values from the cashflow measure for the capacity test for Q4 2019 have significantly changed. Why has this happened?

Callum M posted this 28 November 2019

The values changed because the data tab was previously erroneously displaying results from the Energy stress scenario, regardless of which chart the user had been viewing. They've fixed that now and there's a drop-down box for choosing the scenario.

Will come back to you on the interpretation later this morning.

Cheers,

Callum

Callum M posted this 28 November 2019

Hi Darin

The file attached to the aforementioned thread (https://forum.emi.ea.govt.nz/thread/stress-test-reports-now-on-emi/) has an explanation of how to interpret this (and other) stress test results.

But yes, that's a correct interpretation. If a scenario perturbed a disclosing particpant's cash flow for the coming quarter by -$5, and their last audited accounts had a total cash flow of $100, that (or any figures with the same proportionality) would produce a result of -0.05. In the context of a lower quartile of -0.05, this means that a quarter of disclosing participants had results lower than that and three-quarters had results higher than that.

Cheers,

Callum

 

daringodber posted this 28 November 2019

Great, thanks Callum. That all makes sense to me now. I had read the file but it is still pretty complicated. Personally, I would find some examples (such as my one above) really useful in understanding this.