The four functions of public interest models
We, and others, have made the case before that open financial models is the natural next stage of extractives governance. First, because you don’t know what you’ve got til its modeled, and secondly, the models themselves must be open to allow the cumulative learning that is really important for independent expertise to take hold.
In fact, I would go so far as to say that modeling is the prerequisite for any work on the numbers around extractives. An African colleague asked me the other day if we could train the members of his organisation in how to write a financial analysis of a contract but without using a model – just the written report. A moment’s reflection brought the realisation that no, this was not possible. Any analysis not based on a model would have no validity. It would be like reviewing a book you haven’t read.
But what I’d like to do in this series of blogs is consider the role of financial models in the public interest. Because the more we at OpenOil have got into this area, the more it has become clear that the kinds of models needed are both different to, and have more uses than models traditionally used by industry, and governments (when they do use models).
Models have traditionally served companies and the specialist arms of government who negotiate with them.
Companies run sophisticated analysis of potential profitability before and during their negotiations with governments. The metrics which dominate here are the Expected Monetary Value of a range of different outcomes, Net Present Value and its twin sister the Internal Rate of Return, adjusted to fine-tuned estimates of risk of all kinds – geological prospectivity, fiscal and political stability, and so on.
Oil ministries and national oil companies run many of the same metrics, assessing both the position of the companies and themselves. They may also look at what level of guaranteed income they will get from a project under a range of price and production scenarios, how “front-ended” payments are (what proportion of payments might come early in the lifetime of these generation-long projects), the specific implications for the national oil company if it is going to have a role in the project, and above all the “government take” – what proportion of profits relative to project turnover will come back to the state.
To be sure, these metrics will play a role in the analytical function of public interest modeling. But such analysis is only one of four major functions we see for open models (that is to say, published on the Internet), based on public domain information. Below is a headline summary of these functions, each of will be expanded in a dedicated post.
Public interest models will deal with many of the same questions as traditional modeling, but with an emphasis on ease of access and understanding, and a responsiveness to local issues and attention. To borrow a horrible phrase from IT marketing, public interest models will be “user centric”. So for instance, the falsifiable test of whether a model succeeds or not might be if it could be used to explain the three or four major characteristics of an oil project to a non-specialist audience within 30 minutes, not whether it had modeled all possible variables, or used a sophisticated future pricing scenario which simulates the volatility of the market. In a country where a defined revenue stream is allocated sub-nationally, to a district administration or the communities around the project, these revenue flows might be given major prominence in the model even if their calculation (5% of state dividends) might be considered trivial from a purely technical point of view.
There is a clear pedagogical role for public interest modeling. EITI has achieved tremendous success in process, opening up extractive industries to public debate and legitimising the right to know. But EITI is a means to an end – the systematic understanding of how these industries work, available for public understanding and informed by expertise which is independent of any vested interest, whether companies, governments or international institutions. Modeling is part of that EITI end goal – a big part. Civil society cannot assess questions such as dependency on extractives, public financial planning or whether a particular contract represents a fair deal, without having its own embedded expertise to analyse the many moving parts of finance in an oil or mining project. Public interest modeling can be said to have succeeded when every EITI national secretariat has access to models from their country which they trust and understand, and when at least five people in every EITI country outside business and the government have enough expertise to manipulate and adapt them.
Models based on public domain information have enormous potential to guide advocacy campaigns for more transparency. They are a potent illustration of the value of contract transparency, of course. But beyond that, the many data inputs and estimates needed to make each model run, when they come from public domain, are necessarily imperfect, heterogeneous, and all too often generic. The paradoxical beauty of such a model is that, to exactly the extent that it has wide margins of error from an analytical point of view because of the imperfections of its data inputs, it serves as the basis for a targeted campaign to get better data. Think of this as “keyhole surgery transparency”.
Generic transparency dialogue:
Activists: The government and industry should publish everything.
Government and Industry: Why?
Activists: Because you should! Because it’s the right thing to do!
Keyhole surgery transparency dialogue:
Activists: The government should publish the historic posted prices of crude from this field since the start of production.
Activists: Because it will close a $300 million margin of error in predicting revenues to the government from this oilfield, caused by having to model between two different equally authoritative estimates, which is the best we have right now. But you have the data to close this gap and create greater certainty.
Strengthening Government Capacity
The first and most obvious implication of a model published on the Internet is that it will strengthen the public’s ability to get a handle on projects and contracts. It is natural to assume then that such models might challenge governments, since it will open past negotiations and current management of projects to greater scrutiny.
That assumption is not wholly without foundation. But it is important to understand that public interest models will strengthen government capacity in at least two significant ways – whether they acknowledge it or not.
First, governments have access to better data than is in the public domain. So they can download models and put their own data in, whether or not they publish the results.
Second, the model is equally open to all parts of government. Experience suggests contracts and all related information are often a close hold by line ministries and specialist agencies, and indeed this secrecy within government has significant impact both on capacity to manage and in enabling corruption. But with an open model anyone in the finance ministry, tax authorities, audit agency, investment board, regional governments, prime minister’s office, ministry of the environment – or anyone anywhere in fact – can achieve a basic understanding of the economics of the project, and factor it into their workflows. Inconspicuously if necessary.
The Whole Picture
We need to look at the whole picture when it comes to public interest modeling, because although it will be based on the core concepts of project economics, which have been refined over decades by industry and governments, it will have different emphases – and many additional functions.
And a whole bunch of new constituencies. The role of the public interest model could be defined as offering expertise for the non-expert. As such, we will also need to refine our understanding of the various constituencies interested in these financial aspects of governance.
Civil society has had virtually no access to or capacity to deal with financial modeling, so may be considered a new constituency. But there is vital nuance in the way we should think about the other constituencies.
It will be too simplistic, for example, to talk about how “government” uses financial models. Are we talking two or three experts in the line ministry or national oil company, who may already have complex models based on their specific projects, or much broader circles of civil servants in a wider range of institutions – central bank, finance ministry, tax authorities, audit agencies, regional administrations – who could benefit from independent insight into project economics?
Even in the private sector, public interest models are likely to attract interest among sectors who are not themselves the primary deal makers – the integrated oil companies – whether it is local compliance companies involved in the accounting and legal aspects of the industry or service companies downstream of the main contracts.
There is growing interest and advocacy for open financial modeling to bring about the next stage of transparency. But unless we think through the full potential and implications, we could waste a couple of years producing models exactly like they always have been – for very different audiences and needs. Like the first years of TV, when producers broadcast audio with a static image and called it “television”.