Making a mistake over margins of error

Paul Gunton
Paul Gunton

14 May 2019

I attended the first day of IMO’s Marine Environment Protection Committee (MEPC 74) yesterday (13 May). Most of the time, I am impressed with how IMO committees work. Their chairmen are skilled at controlling the debates and delegates are polite and deferential.

Nations that might be at odds with each other on the international stage can unite behind a common cause in IMO’s debating chamber. Consensus is IMO’s way; it is very rare for a vote to be taken.


But that sense of unity was put under strain on Monday morning, with a number of delegates taking a position that was quite incomprehensible to another group. And incomprehensible to me.

It came at the start of a discussion about air pollution and energy efficiency, with the outcomes of the sixth meeting of MEPC’s Pollution Prevention and Response sub-committee (PPR 6) under review. It had met in February and prepared some draft amendments to MARPOL Annex VI to ensure consistent implementation of the upcoming 0.5% sulphur limit and these are due for approval by MEPC 74 and adoption by MEPC 75.

So far, so ordinary. But then a submission by the International Petroleum Industry Environmental Conservation Association (IPIECA) and the International Bunker Industry Association (IBIA) came up for discussion. It outlines the organisations’ concerns about a change proposed by PPR 6 that could result in compliant fuel being declared non-compliant, thanks to the margin of error in testing procedures for delivered fuel. This was an unintended consequence of an attempt to simplify verification procedures, their paper says.

You can read their paper, MEPC 74/10/11, via IMODOCS, so I won’t quote much from it here, but it argues that, under an ISO standard, for a target set at 0.5% of something, a margin of 0.05% is allowable in reproducing an original measurement; it indicates a ‘95% confidence level’, in statistics-speak.

It gives an example: if a fuel supplier aims to provide compliant fuel with 0.47% sulphur but a later test of the delivered fuel measures it at 0.51%, that is within the ISO-accepted margin of error and should not mean that the fuel is non-compliant.

That made sense to me. But with the amendment proposed by PPR 6, any test result of such a fuel delivery that exceeded 0.5% sulphur would be taken as proof that the fuel does not meet requirements, with no opportunity to make a second test, as is currently allowed. That is the change that PPR 6 has proposed and which IPIECA and IBIA have challenged. Perversely, that margin of error is accepted when making in-service tests.

For a fuel of 0.47% sulphur, they estimate there is a one-in-20 chance that a test will find it is apparently over the limit. For fuels of 0.48% sulphur, the chance of a test giving a result exceeding 0.5% increases to one-in-eight.

This is so obviously unfair to both fuel suppliers and ship operators that I expected delegates to nod sagely, realising that PPR 6 had made a simple mistake that must be reviewed. But no: state after state said they could not support the proposal. Reviewing my notes, an astonishing 34 member states expressed a view, 16 of them supporting IPIECA/IBIA and 18 opposing reopening the discussion.

Each of those 18 was either European or Scandinavian; each of the other 16 was neither European nor Scandinavian; I’ll leave others to speculate how such a definite geographic division of views came to form.

The first delegation to offer its views when the floor was opened to discussion said that it could not support IPIECA and IBIA because “this issue has been truly discussed at PPR 6” and all the others among the 18 who gave a reason all said much the same.

Yet PPR 6 has clearly made a mistake; we all do from time to time. I make them on at least a daily basis, often following a great deal of thought and discussion. But once they come to light, I take any opportunity available to put things right. Are decisions made by IMO sub-committees any different?

One of the larger flag states that supported reviewing the decision clearly shared my frustration. It appealed to those who did not support reviewing PPR 6’s decision. “This delegation has a very hard time understanding why countries would want to prohibit appealing a test result that we know has variability,” its spokesman told the meeting. “I don’t see the problem,” he said. Neither do I and none of those that stood by PPR 6’s decision took to the floor to explain their choice.

What will happen now? This all took place on the first day of the meeting and a working group was to be convened so the conference chair has referred the document for consideration by that group. Even that decision then started a new division in the room: delegations from small nations do not have enough delegates to devote resources to every working group and some urged that a decision be made in the plenary session.

They gained support, with 11 states – by my count – wanting to thrash it out in plenary with just two opposing that proposal. But the chairman stood firm: to debate it in the plenary session would add several days to the meeting, he said, and he was probably right; nearly an hour had been spent just to get this far.

We will learn on Friday what the working group decides. You know what I think they should do, but it could go either way. I certainly do not have anything like 95% confidence in predicting the outcome of this debate.