Introduction
IFC Performance Standard 1 (paragraph 17) requires that the client establish a "management program" that translates the mitigation measures identified in the impact assessment into operational actions. The same requirement is found, with variations in wording, in AfDB Operational Safeguard 1 and in World Bank Environmental and Social Standard 1. From a legal standpoint, the ESMP (Environmental and Social Management Plan) fulfils this role.
From an operational standpoint, it is another story. I have seen ESMPs as bulky as telephone directories, coherent on paper, yet invisible after three months of execution. I have seen others, more streamlined, that remained relevant throughout a five-year construction site. The difference stems neither from the quality of the consultancy, nor from the lender's stringency. It stems from four structural choices that must be settled before the first kick-off meeting.
The four choices in question concern the granularity of measures, the chain of responsibility, the monitoring system and the revision plan. Each can be botched without anyone noticing at the moment of signature. Each is paid for, six months later, in repeated non-compliances and unaddressed complaints. This article explains how to make them correctly.
The ESMP is not one document, it is three documents stacked
The first mistake is to treat the ESMP as a single deliverable. In reality, it must carry three distinct layers of language, addressed to three different readers.
The first layer is that of commitments. It repeats, sometimes word for word, the measures provided for in the ESIA and in the covenants of the financing agreement. These sentences do not change without formal amendment, because it is on them that the lender released the funds. This is the legal layer of the document.
The second layer is that of procedures. Each commitment becomes an operational rule that can be applied without having read the ESIA. A commitment such as "reduce dust emissions towards sensitive areas" is transformed here into "water access tracks and storage areas three times a day in the dry season, record watering in the site log, carry out weekly visual inspection". This is the managerial layer.
The third layer is that of tasks. Each procedure is broken down into named actions, dated, with a verifiable deliverable. Who waters? At what time? With which tanker? Who signs the log? Who records? This is the layer that is missing from almost all the ESMPs I have opened during supervision phase, and it is the one that explains why a technically correct ESMP can produce no result in the field.
Separating these three layers in the document's table of contents, with clear numbering, is the first gesture that changes everything. The site manager looking for what he must do should never have to read the legal layer.
Granularity of measures: between vague and inapplicable
PS1 speaks of "practicable and cost-effective" measures (paragraph 15). The concrete translation of this requirement is less obvious than it appears. A measure that is too vague escapes all control, a measure that is too prescriptive becomes counter-productive as soon as a local constraint imposes a deviation.
The test I use before signing an ESMP matrix is as follows. I take a measure, I pass it to a works supervisor who has not participated in the ESIA, and I ask him what he thinks of the operationalisation of the approach. If he hesitates, the measure is too vague. If he says "I will not be able to do anything else than that, even if reality changes", it is too prescriptive.
The correct wording comes in two stages. First the obligation of result, which does not change: the threshold to be met, the expected level of performance. Then the recommended means, presented explicitly as modifiable subject to documented equivalence. This distinction gives HSE teams the necessary latitude and preserves traceability vis-à-vis the lender.
Chain of responsibility: three names per measure, not a function
PS1 requires that each measure be accompanied by an "allocation of roles, responsibilities and authority" (paragraph 20). This requirement almost always translates into a "responsible" column in the ESMP matrix, filled with generic functions: "HSE Officer", "Project Management", "Contractor".
This wording does not hold up in audit. When a lender arrives on a supervision mission and asks who is responsible for measure 3.4.2, they expect a name, not an organigram. If they receive a function, the conversation inevitably drifts to who was in post at what date, who signed what, and traceability disappears.
The operational rule is that of three names per measure. The one who validates (signatory, generally the project director or the operator's E&S manager). The one who executes (works supervisor, team leader, subcontractor identified by name). The one who verifies (E&S supervisor who must be independent of the producer of the measure). Confusing the executor and the verifier is the most frequent and most criticised failure.
The chain must also provide for a documented escalation. When the verifier identifies a gap that does not resolve, to whom does the alert escalate, within what timeframe, and from what threshold does the information leave the project perimeter? This pathway is what transforms formal monitoring into piloting.
Monitoring system: measurable, proportionate, recipient identified
The monitoring plan is reviewed at each lender supervision mission. PS1 (paragraph 22) calls for indicators enabling the measurement of "the effectiveness of the management program". Three questions structure the answer.
Firstly, which indicators are technically measurable with the means available on the project? Entering a weekly measurement of NO2 in an ESMP when neither the equipment, nor the partner laboratory, nor the sampling protocol are provided produces a phantom indicator that will end up in audit non-compliance. Better a robust qualitative monitoring, actually carried out, than an ambitious quantitative monitoring that remains theoretical.
Secondly, at what frequency? Each type of impact has its own clock. Community grievance handling benefits from being weekly to remain reactive. Physico-chemical monitoring of run-off water can be monthly under normal regime and event-based after each intense rainfall episode. Entering both rhythms in the ESMP avoids mechanically aligning all measures on a quarterly step for reporting convenience.
Thirdly, for whom? Each indicator must have an identified recipient, internal as well as external. This information conditions the form and schedule of reporting, which the ESMP must anticipate to prevent the data producer from rediscovering, three weeks before submission, that they must produce an unplanned dashboard.
Revision plan: keeping the document alive throughout the project
An infrastructure construction site is never stable. The final alignment differs from the alignment studied, subcontractors change hands, seasons impose new constraints, a valid technical modification moves a footprint by a hundred metres. An ESMP that ignores this movement loses its relevance within a few months.
The discipline to install from the initial version rests on two types of revision. An ordinary revision, annual, which integrates monitoring feedback and foreseeable developments. An extraordinary revision, triggered by a significant incident, a substantial project modification, or an explicit request from the lender.
Each revision produces a change table: old wording, new wording, reason, approver, effective date. This table is signed by the project director and transmitted to the lender with the revised version. Without this discipline, the ESMP drifts silently and the version used on site ends up diverging from the version known to the funder. In case of dispute or external review, this divergence can be sufficient to tip a dossier to the unfavourable side.
Conclusion
An effective ESMP is not a document that is perfect at signature, it is a system that absorbs the unforeseen without losing its grounding in the commitments made. The four choices discussed appear technical; they are in reality governance choices. Deciding who decides, who measures, who adjusts, and according to what discipline.
No lender expects an infallible ESMP. All expect a traceable ESMP. The nuance is small in appearance; it is decisive in the relationship that is built project after project between an operator and its funders.
The question to ask before validating the final version is therefore not "is this document complete?" but "will this document survive the first incident that was not foreseen?". If the answer hesitates, one of the four building blocks worked on here is still missing.
Comments
Be the first to react to this article.