Monthly Archives: June 2005

58 – The Hockey Stick again

I have referred previously to the very public debate about the so-called “hockey stick”, a central feature of the most recent report of the Intergovernmental Panel on Climate Change (IPCC) (see PD#6). There have been some interesting developments this year, and a particularly interesting one just recently.

Background: Mann et al. (1998) used a large set of proxy data (mainly measurements of tree ring widths) to estimate temperatures going back centuries. They found that temperatures have been pretty stable since 1400, until they suddenly increased, starting around 1900 (described as the “hockey stick” result). It’s a striking result, and the IPCC featured it very prominently in their last set of publications. (Interestingly, Mann himself was lead author of the chapter that featured his result.)

A geologist (McIntyre) and an economist (McKitrick) tried to reproduce the Mann et al. study, and in the process found errors in the basic data, and a major problem with the statistical method. With errors removed, the hockey stick shape is completely lost, and temperatures in the 1900s seem pretty unremarkable. M&M also found that the model has absolutely no statistical power or significance, with an R2 of 0. There are now three refereed publications by M&M describing their work on this, including one in the journal Geophysical Research Letters, where one of Mann’s papers was published.

Remarkably, Mann et al. have rejected all criticisms of their work. Worse than that, they have refused to provide the full data set and computer code that they used, so that the results could be reproduced independently. The behaviour of Mann and his team has been rather disturbing given that they are meant to be scientists.

Reading through all the material (and there is a lot of it), it is pretty clear that McIntire and McKitrick have thoroughly won the argument.

Mann’s refusal to lie down and concede defeat, even after he has been comprehensively maimed and dismembered, made me think of the wonderful movie, “Monty Python and the Holy Grail”. You have probably seen the scene where King Arthur (M&M) has a sword fight with the Black Knight (Mann). Arthur hacks off one of the Black Knight’s arms, then another, then one leg, then the other, but through it all the Black Knight refuses to concede, and keeps haranguing Arthur to continue fighting.

Arthur: What are you going to do? Bleed on me?

Black Knight: I’m invincible!

Arthur: You’re a looney!

This is such a wonderful description of the hockey stick “debate”. There is Mann, just a head and a bloody torso, madly clinging to the belief that he is still in the fight.

He had appeared determined to just bluff it out and hope to get away with it, but just recently the U.S. House of Representatives Energy and Commerce Committee has gotten involved. It has sent letters to Mann and his co-authors, requiring that their data and statistical code be released, as well as some other interesting related letters to the IPCC and the US National Science Foundation. The letters are publicly available here.

It looks to me like Mann might finally have met his match. I don’t think he will be able to get away with the usual bluff and bluster he uses, but I wouldn’t be surprised if he tries to. It will be fascinating to see whether he does.

In my view, the question of whether the hockey stick result is correct should not be that important in terms of influencing the sorts of actions we undertake. However, the whole saga is important for revealing the amazingly low quality of research processes used by some of the key scientists involved in the IPCC, and the IPCC’s failure to properly audit studies that they put forward as being important and which they hope will have an influence on the expenditure of billions of dollars.

David Pannell, The University of Western Australia

Postscript, 10 July 2005. Email from Glenn Fox: “Here is a modified hockey stick that I made to present to McKitrick and McIntyre at a recent policy workshop held by the Fraser Institute in Toronto. As I explained at the presentation (after McKitrick presented a summary of his work on the hockey stick), as a Canadian and a Hockey player, I was saddened to hear of the loss of the iconic hockey stick as a representation of long-term global temperature. So I went into my workshop and constructed the McKitrick/McIntyre hockey stick so that we could save the icon and be more consistent with the data.”

Further Reading

Web site that describes the debate and provides access to several of the papers:

Steve McIntyre’s blogg, for latest developments in the saga:

The web site of Mann’s defenders:

57 – Herbicide resistance: Does prevention pay?

People often say that preventing a problem is better than curing it or living with it. When you look closely at a particular problem, the old saying is sometimes true, sometimes not. That is exactly the case with herbicide resistance. Past economic analysis has found that prevention usually doesn’t pay, but we’ve recently found that it can do in the important case of glyphosate resistance.

There have been a number of past economic analyses looking at whether it is worth using strategies to delay the onset of herbicide resistance. This studies have tended to conclude that it is not economically attractive to use these strategies pre-emptively (prior to the onset of resistance) largely because the only effective strategy for delaying resistance is not using the herbicide (e.g. Pannell and Zilberman 2001). There is certainly a benefit in preserving a herbicide for use later, but it often comes at the cost of not being able to use the herbicide now, and that cost tends to outweigh the benefit. For most herbicides, there doesn’t seem to be a strategy that will allow more addition applications of a herbicide in future than you have to give up in the short term.

However, at least in Australia, there seems to be an exception in the case of the important non-selective herbicide glyphosate. The frequency of glyphosate resistance genes is low enough in most Australian crop fields for it to be possible to drive those genes to extinction. This could happen if farmers use additional control practices to eliminate any weeds that survive glyphosate application. A particular version of this strategy, known as “double knockdown”, involves a follow-up application of the herbicide paraquat after glyphosate. Biological modelling of this option has been encouraging.

Some colleagues and I set out to analyse the economics of the resistance problem for glyphosate, to see if preservation of the herbicide could be economically attractive (Weersink et al., 2005). There are several complexities behind this question, including the need to spend more in the short run to reduce costs in the long run.

Because every farmer has a different history of herbicide use, we had to think carefully about how to do the study in a way that would be relevant to all farmers. Our approach was to calculate the “break-even period before resistance onset”. If the farmer expects glyphosate resistance to occur quite soon (within say a few years), it is more likely to be worth trying to prevent it. If resistance is not expected for a long time, it is less worthwhile taking preventative action now. The break-even period is the number of years at which the result switches from prevention being worthwhile to not being worthwhile. If resistance is expected to occur within the break-even period, a farmer would benefit from adoption of the resistance-avoiding strategy, even though it is more expensive in the short-term.

The break-even periods we calculated in our economic model varied from three years to 26 years, depending on factors such the extra expense that would occur once the herbicide became resistant, and the interest rate. For realistic assumptions, the break-even period was often around 10 years.

This raises the question of how many years farmers expect to be able to use glyphosate before resistance kicks in. A national survey of 380 Australian grain growers in 2003 found that, on average, grain growers expect that they will get glyphosate resistance in at least one field in 12 years, with 8 percent of growers expecting it in less than five years and 31% in less than 10 years.

So it seems that there are quite a few farmers who expect glyphosate resistance to occure before the break-even period, meaning that they would benefit economically from adopting the double-knockdown strategy. Those farmers with expectations for more rapid onset of resistance are the ones who should adopt it.

On the other hand, it is not worthwhile for all farmers. For farmers with average expectations about the time until resistance onset, it would only be worth investing in prevention in a particular set of circumstances. The farmer would need to have a relatively long planning horizon, a relatively high expected cost of weed control after resistance onset and a relatively low interest rate.

For most herbicides it seems that abstinence is usually the only effective way to prevent resistance development, leading to an unhappy choice between strategies that might be characterised as “just say no” and “accept the inevitable”. For glyphosate resistance in Australia it appears that a third option is sometimes optimal: “short-term pain for long-term gain”. For this third option to be worthwhile in other places as well, there needs to be a weed control option that is not too expensive and is highly effective at killing the weeds that survive glyphosate application.

David Pannell, The University of Western Australia

Further Reading

Weersink, A., Pannell, D.J., and Llewellyn, R.S. (2005). Economics of pre-emptive management to avoid weed resistance to glyphosate in Australia. Crop Protection 24: 659-664. full paper (77K)

Pannell, D.J. and Zilberman, D. (2001). Economic and sociological factors affecting growers’ decision making on herbicide resistance. In: D.L. Shaner and S.B. Powles (eds.) Herbicide Resistance and World Grains, CRC Press, Boca Raton, pp. 251-277. full paper (110K)

56 – Thinking like an economist 18: Constructive roles of economics

This week I consider how economics can contribute constructively in an organisation dealing with agriculture, natural resources or the environment. How can economics particularly help?

The following few observations reflect my experience in applying economics in several organisations dealing with agriculture, natural resources or the environment. In most cases, I am referring to the type of economic modelling that involves representing a fair amount of biological and physical information within the economic model. I call them “bioeconomic” models.

Economics as a tool for integration

Some types of economic model require the economist to work with other disciplines. For example, in building a model to analyse the options for managing a pollution problem, we rely on others to help us understand and quantify the available strategies for pollution abatement, the impacts of those different strategies on the level of pollution, and the the effects of pollution on health, the environment, or businesses.

Economic models of this type provide a framework for pulling together the various disciplines into a coherent whole. They force different disciplines to pitch their contributions at a level that is suitable for the decisions that need to be made, avoiding the common problem of scientists providing information at too fine a scale. Everyone contributing to the model has to operate at a similar level of resolution. I find that working with other disciplines to develop an economic model often forces the disciplines involved to talk to each other in ways they would not otherwise have done. It also clarifies how the different aspect of knowledge should be fitted together to address the problem. Even if the outcome being sought from the decision is not financial, an economic-style framework, applied well, can provide slots for all the relevant information, can help the players see which information is relevant and needed, and ensures that the task is tackled in a holistic way. I don’t see any other approach providing all this.

Asking the right questions

In dealing with managers and other scientists, I have many times found my most valuable role to be in clarifying the questions that we need to address, in order to resolve the issue at hand. This is partly about getting things pitched at the right level of detail, but it is also about having an intuitive feel for the elements of a well-structured decision problem, and recognising what is missing in any particular case.

Dealing with the trade-offs between goals

In agriculture, natural resource management or the environment, it is rarely the case that there is a single high-level objective, such as increasing profit. Usually we are skating around between several objectives, some of which work together and some of which conflict. For example, we might be concerned about risk, resource degradation, downstream impacts, or social impacts, as well as profits. If it is possible to relate those objectives to the available management options, then it is possible to include them in an economic model, which can then be used to tease out the trade-offs (or synergies) between them. If it is not possible to specify the relationships between actions and outcomes, then they can’t be meaningfully considered in the decision,

Broadness through economics

Sometimes economists are justifiably criticised for being too narrow in their analyses. In particular, we often are accused of paying too little attention to distributional consequences and issues of fairness. Nevertheless, a good economic analysis of an issue in agriculture, natural resources or the environment can bring forward a much broader set of considerations than is likely to occur otherwise. It can bring in big-picture considerations, such as external costs from using resources in the proposed way, flow-on effects (e.g. through a farm business, or through the economy), and the distributional consequences of a policy or action (though not whether a particular distribution of benefits is to be preferred).

Protecting the broad public interest

A further big-picture contribution that economics can make is representing the broad interests of the community at large. We remind those making a decision that public money spent in a certain way has to be taken away from other possible uses, so any policy proposal has an “opportunity cost”. We remind people that removal of funds from private individuals or businesses through the tax system means that the funds will not be spent in the ways that would have been most beneficial to those people, as judged by them. For those two reasons, we insist on considering and weighing up the benefits and costs of proposals (if not quantitatively, than at least qualitatively) (see PD#8).

David Pannell, The University of Western Australia

Further Reading

Pannell, D.J. (1996). Lessons from a decade of whole-farm modelling in Western Australia. Review of Agricultural Economics 18: 373-383. full paper (61 K)

Pannell, D.J. (1999). On the estimation of on-farm benefits of agricultural research, Agricultural Systems 61(2): 123-134. full paper (61 K)

Pannell, D.J. (2004). Effectively communicating economics to policy makers. Australian Journal of Agricultural and Resource Economics 48(3): 535-555. full paper from journal (138K pdf) also available via the Journal homepage:

Pannell, D.J. (2004). Is economics hard hearted? Pannell Discussions, No. 8, 12 July 2004,


55 – Targeting perennials to balance profit, water yields and salt loads

“Until recently, conventional thinking on salinity was fairly simple – more perennial plants, less salinity,” says Dr Tom Nordblom of CRC Salinity and New South Wales Department of Primary Industries. Now modelling by Tom and his colleagues Iain Hume, Andrew Bathgate and Michael Reynolds has shown that the reality can be far more complex.

In the slopes region of New South Wales, the planting of perennial vegetation is promoted as part of the strategy for salinity management. However, recent modelling has highlighted the need for careful economic and hyrdological analysis to understand where perennials should be established in this region. Poorly sited perennials can result in major economic costs to landholders, major losses of freshwater additions to rivers, and can even make salinity concentrations in the rivers worse. On the other hand, well located perennials can substantially reduce salt concentrations without excessive costs in terms of farm production or river flows.

Modelling by Tom Nordblom and his colleagues allows us to estimate the trade-offs among profitability, water yield and salt concentrations by integrating information on land use, plant water use, rainfall, soils, groundwater salinity, catchment hydrogeology, and economics.

Contributions to stream water flow and salt load differ widely among sub-catchments depending on groundwater salinity, soils and current land use. For example, the 80 sub-catchments of upper Little River were classed into six groups according to their soils and contributions of water yield and salt load to the river (Figure 1).

Figure 1. Upper Little River catchment, divided into 80 sub-catchments


Red indicates sub-catchments that deliver flows of high salt concentration, dark blue sub-catchments are relatively fresh and green sub-catchments are in between. A range of water-yield and salt-load targets can be met in the future by changing land use now. High water yields and high profits are associated with cleared land under cropping, while forests are at the lower end of the spectrum for these outputs. Improved perennial pastures are in between.

The minimum cost of attaining a specific set of target varies widely, depending on the targets. This is illustrated in Figure 2, which shows results for a mini-catchment of three sub-catchments (one red, one green and one dark blue) having a total of 4700 ha.

Figure 2. Opportunity costs for shifts from current land use (yellow dot) to the most profitable land uses which can deliver target future levels of water yields and salt loads in the river.


For comparison, the average water yield and salt load levels under current land use are shown by the yellow dot. The height of each bar shows the lost farm profit compared to current land use. The higher the bar, the lower is the profitability of agriculture in that scenario.

In this example, current flows deliver a stream salt concentration of 300 parts per million (ppm). Changes in land use offer a range of 200 to 500 ppm in the future (Figure 2).

Three example targets (A, B and C) illustrate trade-offs between farm economics, downstream water volumes and salt loads.

Target A is attainable by planting perennials, but, compared to the current set of land uses, it incurs over $6m in cost (lost profits), lowers annual stream flow by 1500 ML and increases stream salinity from 300 to 500 ppm: a lose, lose, lose option.

Target B halves current salt load, costs $0.4m, reduces stream flow by 500 ML, and improves stream salinity to 200 ppm. Target C offers the same improvement in water quality as B but at minimal cost and with little loss in stream flow.

Judging the “best” target is a question of weighing up the agricultural costs against downstream demands for water and water quality. This framework presents to decision makers the full set of tradeoffs on the supply side. Downstream demands for water and water quality by towns, irrigators and for environmental flows comprise the other side of the question. In regions such as the Little River catchment, where flows of fresh water into rivers are significant, a capacity to use such a framework is crucial if resource managers are to avoid the real risk of making river salinity worse while also causing major losses of profits and water flow.

It is important to appreciate that this analysis is specific to river salinity. The issues and the appropriate strategies are likely to be different for management of salinity threats to agricultural land, infrastructure and terrestrial biodiversity.

David Pannell, The University of Western Australia

Further Reading

Nordblom, T., Hume, I., Bathgate, A., Hean, R., & Reynolds, M. 2005. Towards a market: geophysical-bioeconomic targeting of plant-based land use change for management of stream water yield and salinity. Invited paper, 2nd Economics and Environment Network (EEN) National Workshop, ANU, Canberra, ACT, May 5 – 6, posted at