Solutions
Industries
Regulations
Resources
Company
Key Takeaways
A friend of mine works at a company that spent a fairly serious amount of money preparing for CSRD. They hired people, brought in consultants, started organising sustainability data and generally did what companies do when a large European regulation appears on the horizon and everyone tells them they need to get ready.
Then came the Omnibus.
The scope of CSRD was substantially reduced, timelines moved and parts of the sustainability reporting architecture were simplified. Suddenly, companies that had spent the previous couple of years preparing for mandatory reporting were looking at work they might no longer be legally required to do. The fairly understandable reaction was some version of: what exactly did we spend all that money for?
I find this interesting because the instinctive answer is to decide, with the benefit of hindsight, whether the company “wasted” its money. If it ultimately doesn’t have to comply, perhaps it should have waited. If it had waited and the regulation had remained unchanged, perhaps we would now be criticising management for being completely unprepared.
Neither is a particularly useful way of evaluating the original decision. Businesses don’t get to make decisions with knowledge of the regulation that will exist three years later. They have to make them using the probabilities, deadlines, costs and consequences visible at the time.
Once we think about it that way, regulatory compliance starts looking surprisingly similar to any other investment made under uncertainty.
Imagine that in 2023 a company believes there is an 80% probability that it will have to comply with CSRD in roughly its existing form. Preparing early will cost €500,000. Waiting will save that money if the regulation disappears, but if the requirement remains, late preparation could cost €1 million because consultants are scarce, systems have to be implemented quickly, data has not been collected historically and the organisation has very little time to figure out what it is doing.
There is also a risk that being unprepared creates consequences beyond the direct implementation cost. Reporting can be delayed, assurance can become painful, employees can spend months firefighting and management can discover rather late that the company does not actually possess some of the data it assumed it did.
Under those conditions, spending €500,000 early can be a perfectly sensible decision even if, several years later, the regulation changes and some of that work turns out not to have been necessary.
The decision can be good and the outcome can still be unfortunate. We understand this perfectly well in most other areas of business. A company can hedge a currency exposure and later discover that the currency moved in its favour. We don’t conclude that the hedge was idiotic merely because, in retrospect, doing nothing would have made more money. We ask whether the hedge made sense given the risk the company faced when the decision was made.
Regulatory preparation deserves the same treatment.
There is another side to this story, though. Saying that companies made rational decisions under uncertainty doesn’t mean the uncertainty itself was costless.
If thousands of companies hire people, buy software, restructure data systems, engage consultants and redesign processes because a regulation requires them to do so, those are real economic resources. If the regulation is later narrowed because policymakers conclude that the original architecture imposed too much burden, some portion of that investment becomes stranded.
This is one of the reasons regulatory stability matters economically. A government can change a law with a vote and a publication in an official journal. A company cannot reverse three years of organisational investment quite so neatly.
People were hired. Contracts were signed. Technology was purchased. Data programmes were built. Suppliers were asked to provide information. Internal processes changed. Entire consulting and technology markets formed around the expected requirement.
This doesn’t mean a regulator should refuse to correct a bad rule simply because companies have already spent money complying with it. That would be an even stranger outcome. If the regulation is too expensive relative to the benefit it creates, sunk private investment is not a good reason to preserve it forever.
But it does mean that changing regulation has a cost that isn’t captured simply by comparing the old legal text with the new one.
Regulatory uncertainty has a balance sheet, even if nobody reports it.
The European Commission’s simplification agenda is unusually explicit about administrative cost. The sustainability changes under Omnibus I are estimated by the Commission to save businesses billions of euros annually.
That is important because it acknowledges something that can occasionally disappear from sustainability-policy discussions: the cost of producing information is itself economically relevant.
CSRD was built around a reasonable theory. Markets do not always possess enough comparable information about companies’ environmental and social impacts, risks and opportunities. If investors, lenders, customers and other stakeholders receive better information, they can incorporate it into decisions. Better information can therefore influence where capital goes, which suppliers customers choose and how companies manage sustainability-related risks.
This is a different market failure from the one addressed by the EU ETS. ETS says that emitting carbon is too cheap because the environmental cost isn’t fully reflected in the market price. Sustainability reporting says, in effect, that important information is missing or inconsistent, so market participants cannot make fully informed decisions.
But information has a cost to produce. There is therefore some theoretical point at which asking for one more disclosure costs more than the decision-making value that disclosure creates.
Finding that point is obviously difficult. Nobody can tell you precisely that datapoint number 783 in an ESRS disclosure is worth €14.62 to European capital markets. But the inability to calculate the optimum perfectly doesn’t make the trade-off disappear.
The Omnibus is significant because Europe has effectively acknowledged that some parts of its sustainability architecture had moved too far along that cost curve.
There is a temptation, particularly after a large regulatory rollback, to rewrite the entire history as though the original policy was obviously absurd. I don’t think that is particularly convincing either.
Before CSRD, sustainability reporting was fragmented, methodologies varied widely and companies had considerable discretion over what they reported. Investors and customers increasingly wanted sustainability information, while large companies were asking suppliers for data in dozens of different formats. Creating greater standardisation and comparability had a perfectly coherent economic rationale.
The difficult part was determining how far that standardisation should go, how many companies should be subject to mandatory reporting, how much detail they should provide and how much information large companies should effectively be able to demand from smaller companies throughout their value chains.
Those are questions of proportionality rather than ideology.
A reporting requirement can produce useful information and still be too expensive. A regulation can solve a real problem and still solve it inefficiently. A policy can be directionally correct while its first implementation is too broad. None of these things are contradictions.
In fact, if we expect governments to regulate complex new areas such as climate transition, some degree of correction is probably inevitable. The alternative is pretending that policymakers can design a perfect system on the first attempt and then refusing to change it when reality produces inconvenient evidence.
The more important question is how expensive the learning process becomes.
There is another consequence of regulatory revision that I think matters beyond CSRD itself. Companies learn.
If a regulator announces a requirement and businesses spend heavily preparing for it, only to see the requirement substantially changed before it fully bites, the next regulatory announcement may not produce the same response. Management teams may wait longer. Procurement may delay buying software. Boards may be less willing to approve large implementation budgets until the legal position becomes clearer.
That reaction can be entirely rational.
It can also create a problem for policymakers. Regulation often relies on companies beginning to invest before the final compliance date. Renewable-energy targets are useful partly because developers believe future demand will exist. Product regulations influence engineering decisions years before a product reaches the market. Carbon rules influence investments in assets with operating lives measured in decades.
If businesses begin applying a large “regulatory credibility discount” to every new European target, policy becomes less effective even when the eventual regulation remains unchanged.
This is why I think regulatory certainty needs a slightly more nuanced definition than simply asking whether Europe is committed to decarbonisation. The direction of travel can be extremely clear while the individual instruments remain uncertain.
Europe can remain committed to reducing emissions while changing how CSRD works. It can remain committed to carbon pricing while adjusting ETS rules. It can remain committed to renewable fuels while changing feedstock treatment or compliance mechanics. It can remain committed to product sustainability while reconsidering exactly which information companies must provide.
For businesses, certainty of direction and certainty of instrument are two different things.
There is no satisfying universal answer, but I think the mistake is treating compliance preparation as binary. Companies often behave as though the choices are either “do nothing until absolutely required” or “build the complete compliance machine immediately.”
A better approach is to think about reversibility.
Some investments remain useful even if the regulation changes. Understanding your major emissions sources, improving supplier master data, knowing where critical sustainability information sits inside the organisation and establishing basic data governance can have value across several regulatory regimes. These are relatively robust investments because they solve underlying information problems rather than merely satisfying one paragraph of one regulation.
Other investments are highly regulation-specific. Building elaborate workflows around a disclosure requirement whose final scope is uncertain can create much more stranded cost if the rule changes. The closer an investment is to the exact mechanics of an unsettled regulation, the more valuable flexibility becomes.
That suggests a fairly ordinary business principle that sustainability teams sometimes forget because a regulation is involved: preserve optionality where uncertainty is high.
Build the capabilities you are likely to need across multiple scenarios. Delay irreversible expenditure that only makes sense under one narrow interpretation when you reasonably can. Understand the cost of being late if you wait. Then make the decision based on the probability-weighted downside rather than on whether somebody on LinkedIn says the regulation is definitely happening or definitely dead.
Regulatory compliance is still capital allocation.
Software built around regulation has the same problem. If every workflow is hard-coded around today’s version of a rule, every regulatory amendment becomes an expensive redevelopment project. The customer then pays twice: first to implement the original compliance requirement and again to change the system when policymakers change their minds.
A better architecture assumes that regulations will move. Data requirements change, thresholds change, methodologies evolve and occasionally entire obligations disappear. The underlying commercial and sustainability data should therefore be more durable than the regulatory workflows using it.
This is another reason I am increasingly interested in the difference between sustainability data and sustainability reporting. A supplier’s product carbon footprint can remain commercially useful even if one reporting obligation changes. Knowing whether a material is certified can matter to a transaction independently of whether that information appears in an annual report. Understanding the carbon intensity of a production route can affect procurement, pricing or investment even if nobody requires the company to publish it.
The closer sustainability information gets to an actual economic decision, the less vulnerable its value is to the fortunes of one reporting regulation.
There is a very understandable temptation after CSRD and Omnibus to conclude that companies should stop spending money until regulators have completely finished changing the rules. Unfortunately, regulators are never completely finished changing the rules.
The energy transition is going to involve decades of technological change, political negotiation, industrial competition and learning. Some regulations will tighten, some will simplify and some ideas that currently look inevitable will probably disappear entirely. Waiting for perfect certainty is therefore another way of deciding never to move.
The better lesson is to become more sophisticated about regulatory risk. Ask how durable the policy objective is, how stable the particular instrument appears to be, what happens if you prepare too early, what happens if you prepare too late and how much of the investment remains useful under alternative scenarios.
That is a much more useful conversation than deciding afterwards who was stupid.
And it leads to another regulation where the economic consequence is becoming considerably harder to treat as a sustainability-reporting problem. With CBAM, the carbon number attached to a product begins influencing what it costs to bring that product into Europe.
At that point, the question isn’t really “are you green?” anymore.
It is: what does your carbon cost?
That is where I want to go next.
