The UN is wrong

The UN is wrong

The UN is wrong if they establish a standardised chemical hazards classification database. They would be taking a fatally flawed strategic path. The correct strategy is outlined below.


It is necessary to classify the hazards of all substances being used in workplaces so Safety Data Sheets (SDSs) can be written and recommended protections can be implemented.

The problem is easy to understand. Different experts, being humans with varying education and experience, classify hazards of the same substance differently. Therefore, some workers are at more risk than others depending on which classification they see.

To address this problem, UN’s GHS Sub-Committee[1] is considering creation of a database of “standardised” hazard classifications. They have just done an international pilot study to produce consensus[2] classifications for three substances to discover it takes considerable time and expertise to analyse each one.


But this is blindingly irrelevant because worldwide every supplier of chemicals is legally responsible for classifying the hazards of every chemical they supply.

They cannot legally rely on an independent third party. The supplier is responsible.

Lawyers and governments do not write legislation in a way which permits anyone to wriggle off the hook.

A contracted scientist who classifies a substance incorrectly for a supplier will eventually be held responsible in court  - jointly and severally with the supplier  -  if the worst occurs. That is why they buy professional indemnity insurance.

Variations on purity

There are frequently different grades of the same chemicals used in industry. From 100% purity down to some lower percentage which is cheaper as fewer impurities are removed.

Substance properties vary depending on purity. The GHS requires that properties considered during hazard classification must include the effects of any impurities.

This implies the “same” substance may indeed have varying hazards depending on type and quantity of impurities.

Future danger #1

Assume the UN database of hazard classifications is established.

Look ahead a few years and smaller suppliers which cannot do their own research will take short-cuts and adopt the UN classifications of a substance irrespective of the purity they are supplying.

This is a real danger for workers.

Future danger #2

Look ahead a few decades. Research has delivered new information about a substance. Fortunately, the UN has conducted its due diligence and by consensus has updated the hazards classification database accordingly.

Also fortunately, the UN knows the identity of people who registered to use the harmonised or standardised classifications. It therefore notifies everyone on its list.

Unfortunately, because tens of thousands of people are registered and we are considering an event a few decades hence, there will be many thousands of job-changers, retirees and deceased registrants who inadvertently block notification to particular suppliers.

Future danger #3

Look ahead half a century or more. Now consider all the SDSs which were originally based on the UN substance classifications and have been rolled over every few years without the (obviously small) suppliers bothering to genuinely refresh them.

There will be a number of SDSs from different suppliers for the same substance, possibly with different degrees of purity, with differences possibly due to Future danger #2, but in any case with conflicting classifications. Workers reading one set will implement different protections than workers reading another.

Some unscrupulous suppliers may even be tempted to pick and choose classifications to suit their marketing requirements.

SDSs with incorrect classifications will persist for centuries possibly indefinitely on the basis of an ancestor SDS attributing those classifications to the “authoritative” UN classifications database back in time.


The proposed third party UN hazard classifications database will be expensive to build and especially so to maintain.

In the longer term it will fail because the UN does not have primary vested interest in its success. The UN does not use the data itself.

If it can be portrayed in future as a failure the database will eventually lose funding to other UN priorities. Not only will the cost have been high but it will not have a lasting benefit. It is therefore poor economics.

What is really happening?

Large companies do the right thing. Otherwise their Directors will be liable. They would usually not risk relying on a third party classification database ahead of their own research.

On the other hand, small companies usually do rely on the research of others.

All the dangers outlined above are occurring now. They are not waiting for any proposed UN database to suddenly begin.

At best the proposed database will change nothing.

At worst it will entrench the current poor situation forever. Perhaps spreading it even deeper in the industry to larger companies. Why?

The real problem is data

Many large companies keep their data to themselves because they did their own expensive research.

Small companies are forced to rely on data they can glean from supplier SDSs and published databases.

Both these sources have problems. Data from supplier SDSs may be out-of-date, somewhat obfuscated or even missing. Existing published third party databases suffer from exactly the same issues as predicted here for the proposed UN database.

Or the real problem is regulatory dis-harmony

Different jurisdictions have different regulations. This means the same properties of a substance may attract a different hazard classification in different jurisdictions. This is a political problem which needs a political solution. 

While a UN list of standardised or agreed or consensus GHS classifications might ignore regulatory differences, suppliers in (or to) those jurisdictions would not be able to use the standardised classifications. 

Our point here is that standardised classifications cannot solve a political or regulatory dis-harmony problem. 

The correct strategy

The correct strategy recognises that there is only one entity with a real vested interest in keeping substance property data accurate and staying on top of emerging research. That is the substance manufacturer.

The manufacturer owns, uses and relies on that data.

There is only one entity with responsibility for hazards classification. The law around the world makes clear it is the substance supplier.

Any IT solution which does not recognise and address these two immutable truths is fatally flawed.

The correct strategy must hold these two truths to be self-evident! Plus some guiding principles[3] identified by the GHS Sub-Committee.

One strategic solution

There is one solution which has already been developed with a superset of the GHS Sub-Committee’s Guiding principles.

That solution is a database which embodies the efficiencies of not duplicating classification effort and can scale to the size of Instagram or Facebook. It also satisfies requirements of all stakeholders from large chemical companies, formulators, distributors, small suppliers and workplaces.

We have designed and built the free open source SharedSDS accordingly and welcome any support and collaboration.


[1] The UN has a Sub-Committee of Experts on the Globally Harmonized System of Classification and Labelling of Chemicals. This is widely known as the GHS Sub-Committee.

[2] There is nothing wrong with consensus. A second opinion is extremely valuable. Like peer-review, worth its weight in gold. But it means the classification is done at least twice making the process more expensive and time consuming. See also the agenda item proposed for the GHS Sub-Committee meeting 10–12 July 2017 in Geneva which describes the Pilot Project to analyse three chemicals:
Additionally the Pilot Project was co-ordinated by the OECD. This is the OECD report:

[3] Guiding principles. The GHS Sub-Committee specified, inter alia, for their proposed hazard classifications database, “The source of the information must also be electronically available, and publicly accessible.” See Annex III on page 18 here: 

Currently unrated