And why they are wrong
Author’s note: In this issue of Decision Point we have discussed the application of decision science and making more of models (pages 4,5), structured decision making (pages 6-11) and adaptive management (pages 12,13). This is the core territory of our two centres. While the advantages of this approach to conservation are spelt out in multiple ways throughout this and every issue of Decision Point, it’s important to note that many people still have objections to the use of decision science for a variety of reasons. Back in 2009 I wrote an editorial on this very topic for one of the early issues of Decision Point and it seems timely to reproduce it here. The text is largely unaltered from then, and the Stefan Hajkowicz (2008) reference on NRM expenditure is still relevant: we’re still spending billions of dollars on the environment in a largely ad hoc, unaccountable and opaque fashion. While systematic spatial conservation planning tools are now the norm, simpler non-spatial tools for prioritisation, like cost-effectiveness, are only just starting to be used. Further, the recent Senate Inquiry into the Effectiveness of threatened species and ecological communities’ protection in Australia uncovered people expressing discomfort (or disquiet) with using rational repeatable processes for allocating government funds. So it’s a good time to face the doubters again. Hugh Possingham, Director, Environmental Decision Group
Since 1994 I have given over 300 seminars to every manner of audience on how decision science can inform environmental management. In that time I’ve received a wide range of arguments about why decision theory tools should not be applied to conservation problems. Sometimes people reject the general message that anything could be wrong with current decision making. Sometimes they take exception to the details in specific case studies. Sometimes they just get a little angry.
I’d like to take a moment and review the most common objections and suggest that they are wrong. But I’m going to go further and say that natural resource management in Australia must embrace the tools of decision science, like every other rational profession, from medicine to engineering (and it needs to happen sooner than later).
It’s important to dispel the myths surrounding these objections. To ignore the value of environmental decision science is to repeat the many mistakes that we have made in the last two decades. As a nation we’re failing hopelessly to secure our most precious and unique natural asset – Australia’s biodiversity. Since 1990, the Australian federal government has announced seven major natural resource programs collectively worth $6.51 billion (Hajkowicz, 2009). In almost every case the allocation decisions continue to be ad hoc and opaque (despite a lot of excellent advice, see ‘Why INFFER’).
So, here are the five most common objections with their strengths and weaknesses.
Objection 1: It’s based on models
Well, of course, this is correct; all models of everything are wrong. The only perfect model for something is the thing itself, and then it ceases to be a model. However models are our only way of predicting the future and if you are a manager you must be predicting the future as a consequence of actions – otherwise you could never take an action. Hence, by definition, every manager, indeed every human being, is a modeller – it is just that most don’t use maths.
Furthermore there is confusion between what is a model, what is a problem and what is an algorithm. Much of decision theory is about mathematically defining a problem with objectives and constraints. This is the translation of human hopes, dreams and fears into maths, another language. It is not classical ecological modelling. Often there are some more typical ecological models (that either predict in space and time) that lie inside the problem definition, and they should be scrutinised with care. Finally we use algorithms, not models, to find good solutions to the problem (not the model). If one uses mathematically credible algorithms (not scoring systems or overlays of GIS maps) then the solutions are likely to be close to correct.
(See Decision Point #22 for a more detailed discussion on this point.)
“The principles of decision theory are founded in economics and it is economics that has made a mess of the world. Decision science is the tool of the Devil.”
You can kill someone with a hammer but it doesn’t mean that hammers are bad tools. The tools we use are largely from a branch of mathematics called operations research, designed to solve mathematically well-defined problems. These tools are used by engineers, mathematicians and economists (mathematicians who know some big complex words).
The tools of decision science drive most small scale (micro-economic) decisions from how to supply troops in a battle, to oil refineries ordering crude supplies, to airline companies devising their plane schedules.
One aspect of decision science that really irks some people (and is very much an economic way of thinking) is that we generally need to give every value a numerical quantity. Typically economists deal mainly with money, but environmental values often deal with things that at first glance seem harder to quantify – threatened species, ecosystem services and social values.
However, with thought, most can be quantified, albeit with some uncertainty (see point 5) and they do not have to be turned into dollars and cents for them to be used in decision science. One algorithm that provides good solutions to many conservation problems involves ranking actions in terms of their cost-effectiveness – that is their biodiversity benefit divided by their cost (in dollars). This simply requires an integrated and numerical measure of the biodiversity benefits of an action.
Objection 3: It’s too cumbersome
In cases where decisions are small and once-off, you are correct, spending money on using or developing a decision theory tool is like cracking a nut with a steam roller. However tools are becoming easier and easier to use, and many environmental management problems involve millions of dollars. In most cases the use of an approach or tool will save money and biodiversity. Or, more uncharitably, if you are not using the most appropriate decision theory tools you may well be consigning some biodiversity to extinction!
“I’d describe them all as myths that the nation can no longer afford to accept.”
Developing decision theory tools might slow some projects down. In many cases, however, the focus provided by formulating and solving an explicit problem speeds things up. For example consider the project prioritisation protocol (PPP) that the New Zealand environment department has developed and used in partnership with the Environmental Decision Group (Joseph et al, 2009). In less than three years they have not only developed action plans for over 600 species, they have costed and prioritised them – all because the decision-making framework made recovery planning more focussed. Australian agencies have been developing recovery plans for two decades at huge expense: many remain unfinished or incomplete.
See Decision Point #29 for a description of how PPP works.
Objection 4: It’s a black box
You are correct, which is why we always say that these approaches and tools inform decisions rather than make decisions. There are invariably considerations that cannot be accommodated in the formulation of a complex socio-ecological problem – so we need some wiggle room.
As far as black-boxes are concerned – so what? What percentage of the public know how planes fly, or hand calculators calculate, or microwaves cook? Although, if you want to look a bit more closely at the mechanism inside a decision-tool black box, you might be surprised at how simple they are. Many decision-making tools are much simpler than people think if they are explained well and people take some time to work with them. (Consider for example our ‘Marxan out of the box’ discussion in Decision Point #62)
Objection 5: There’s too much uncertainty
Uncertainty and risk are rife in ecological systems – and we face it far more than, say, an engineer (but no more than economists or doctors). Fortunately the mathematicians have many ways of rigorously accounting for uncertainty in decision-making – indeed if a system has a lot of randomness then it is even more important to take a formal quantitative decision-making approach. Furthermore, the ideas of ‘active adaptive management’ (see Decision Point #17) and ‘value of information theory’ (see Decision Point #67) enables us to include the benefits of learning in prioritising our actions.
Some people think that decision theory tools can only deliver perfect but fragile solutions – however there are many algorithms for obtaining robust solutions, solutions that deliver reasonable results under a lot of randomness and uncertainty.
So there are ‘my’ five objections to decision science (they are really classes or streams of objections rather than specific snipes), and I’d describe them all as myths that the nation can no longer afford to accept. After peddling the decision-theory message for 20 years, I think the message is unanimous and unambiguous – it is now the responsibility of managers and policy-makers to make use of the best available science and tools.
Hajkowicz SA. (2009) The Evolution of Australia’s Natural Resource Management Programs: Towards improved targeting and evaluation of investments. Land Use Policy 26: 471-478. doi:10.1016/j.landusepol.2008.06.004
Joseph LN, Maloney RF & Possingham HP (2009) Optimal Allocation of Resources among Threatened Species: a Project Prioritization Protocol. Conservation Biology 23:328-338.