Cost-benefit analysis of pest management

The next stage of economic analysis is to understand how pest management strategies influence economic damage and to conduct a cost-benefit analysis (CBA) of control efforts. The benefits of pest management are the damages that are averted (i.e., changes in economic surplus) as a result of various levels of management effort. These benefits are then compared to the cost of labour and other resources (e.g., pesticides, capital equipment) used in the control effort. Frequently analysis is restricted to a few management scenarios with the project yielding the highest net present value being recommended. However, in some instances, optimal control methods may be applied which seek to discover the control strategies that maximise the overall net benefit. Optimal control methods typically employ dynamic programming techniques and can be quite sophisticated (e.g., Sims et al. Reference Sims, Aadland and Finnoff2010).

A major challenge with the CBA of control is establishing a credible linkage between management interventions and the change in the flow of economic surplus from the forest (Holmes et al. Reference Holmes, Prestemon. and Abt2008). As a result of significant time lags between intervention and outcomes, confounding influences are almost always present. Proper empirical examination can be data intensive and requires statistical rigour. In the absence of empirical evidence, researchers can, and often do, make simplifying assumptions. While these assumptions may have their roots in plausible bio-physical or economic theories, the lack of empirical support serves to increase the uncertainty around the benefits of control.

There are several ways to deal with uncertainty. The most common way to model uncertainty is as risk, which involves assigning probabilities to the possible outcomes of pest management. In fact, the benefits of management interventions are often associated with their reduction in the probability of an adverse outcome (Holmes et al. Reference Holmes, Prestemon. and Abt2008), causing an increase in the expected economic surplus. However, uncertainty extends beyond the physical impacts of control measures on a forest estate: forest products prices and other economic parameters are volatile over time. Monte Carlo simulation is a useful tool to deal with multiple sources of uncertainty.

Cost effectiveness and threshold analysis

From an economic perspective, in an ideal world CBA should be undertaken to guide the choice of pest management programmes. In reality, a full CBA rarely occurs. Instead a programme is chosen based on a variety of goals or is simply based on historical precedent and managers are tasked with implementing the programme as effectively as possible. The target may be a given volume of timber saved, a percentage of infested trees, lowering defoliation levels in order to keep the forest green, or the population size of pests. Under such constraints, there still is a role for economic analysis. Analysis may be undertaken to achieve the programme objective in a lowest cost fashion, termed cost-effectiveness analysis (CEA). One of the computational advantages of CEA over CBA is that CEA does not require an explicit quantification of hard-to-measure programme benefits (Bicknell Reference Bicknell1993). Alternatively, the amount of management effort may be fixed and economic analysis is conducted to uncover the threshold level of pest activity that would justify the treatment. Threshold analysis is common in pest management in agriculture but has also been done in a forestry setting (Fox et al. Reference Fox, Beke, Hopkin and McKenney1997).

Economic impact analysis

Because the forest sector is a dominant component of several regional economies across Canada, policy makers are frequently concerned about the indirect or spill-over effects of the forest sector on other sectors within the region. Economic impact analysis (EIA) is sometimes employed to quantify the comprehensive damage of forest disturbances on a regional economy or the broader macro economy. Such impacts could be inter-industrial, and both forward (downstream industries) and backward (upstream industries) impacts are included. Input-output (IO) models are the most commonly used EIA method (Davis Reference Davis1990). However, IO models are limited by their reliance on linear, static relationships (Lewis Reference Lewis1988). Furthermore, their use in the calculation of the economic damage associated with forest pests is upwardly biased, leading to an overstatement of the benefits of control (Stabler et al. Reference Stabler, van Kooten and Meyer1988; Hamilton et al. Reference Hamilton, Whittlesey, Robison and Ellis1991). Regional dynamic computable equilibrium (CGE) models are a means of overcoming the limitations of IO analysis. Nonetheless, impacts can be very sensitive to the assumptions used in CGE models, particularly with regard to the mobility and opportunity cost of factors.

Spruce budworm

The first CBA of SBW in New Brunswick, Canada was conducted in 1965 by D.A. Wilson of the Canadian Forest Service. Wilson’s CBA supported a Federal-Provincial cost-sharing agreement for a spraying programme in 1965–1966 (Marshall Reference Marshall1975). This was followed up by a study by Marshall (Reference Marshall1975), who also concluded that the benefits of spraying exceeded costs. However, in calculating the benefits Marshall made several questionable, simplifying assumptions. Most notably, he assumed that there were no alternative uses for inputs (labour, fuel, capital) used in forest production and consequently applied the saved valued-added in the forest sector as the benefit of control. This effectively assumes inputs cost nothing (e.g., labour is willing to work for free; manufacturing facilities can be established, maintained and run at no cost), which is not a reasonable assumption in our view.

Significant effort was then invested into understanding the timber supply impacts of the SBW and the influence of forest management, and this effort resulted in the SBW decision support system (SBW DSS) (MacLean et al. Reference MacLean, Erdle, MacKinnon, Porter, Beaton and Cormier2001). The SBW DSS has been a key input in recent economic analysis estimating both market and non-market benefits and costs of controlling hypothetical SBW outbreaks. Slaney et al. (Reference Slaney, Lantz and MacLean2009) conducted an innovative study that examined the costs and benefits of protecting carbon through pest management (i.e., spraying the biological insecticide Bacillus thuringiensis Berliner (Bacillaceae)). The main feature of this study was to integrate carbon sequestration potential and carbon credit benefits into the SBW DSS. By doing so, forest managers and policy makers could then design and assess management programmes based on objectives not limited to just timber value maximisation. Twenty-four alternative scenarios were examined in two forest management units located in New Brunswick and Saskatchewan, respectively. The scenarios included two outbreak severities, three protection frequencies, and four protection area sizes. The study concluded that large-scale protection programmes were worthwhile, particularly in New Brunswick.

The same authors then conducted a similar study (Slaney et al. Reference Slaney, Lantz and MacLean2010) that incorporated timber values. Twelve SBW protection strategies were assessed on the same forest management areas: Crown License 1 in New Brunswick and the Prince Albert Forest Management Area in Saskatchewan. This study also supported intensive protections measures; however, it also revealed that smaller programmes tended to yield higher benefit-cost ratios (BCR) and were more appropriate when timber prices are low.

In a unique non-market valuation study, Chang et al. (Reference Chang, Lantz, Hennigar and MacLean2011) used contingent valuation to estimate the social benefits of SBW and FTC control also in Saskatchewan and New Brunswick. The results of their survey suggested that social benefits can be large, ranging between CAD$14.3–32.4 million per year for SBW and CAD$7.9–22.0 million for FTC. This study, along with an advanced SBW DSS model (Hennigar et al. Reference Hennigar, Erdle, Gullison and MacLean2013), was then used by Chang et al. (Reference Chang, Lantz and MacLean2012a) to conduct another CBA of controlling SBW in New Brunswick. The CBA evaluated six scenarios, which included three proposed protection levels and two unprotected outbreak severities. For each scenario, the net present value and BCR of market values and non-market values were calculated and presented. By including non-market values, CBA metrics (BCR and NPV) improved and additional control tended to be supported relative to the situations where market values were only considered.

Finally, Chang et al. (Reference Chang, Lantz, Hennigar and MacLean2012b) also using the SBW DSS and a dynamic CGE model conducted an EIA of future SBW outbreaks on Crown forest land in New Brunswick. This study projected the impacts over the 2012–2041 period and found that under an uncontrolled moderate SBW outbreak, total output in the New Brunswick economy would decline by 3.3 billion Canadian dollars under a moderate outbreak and 4.7 billion Canadian dollars under a severe outbreak. Aerial spraying scenarios were found to mitigate these impacts by up to 66%, and harvest scheduling adjustments and salvage were found to mitigate impacts further by up to 18%. The authors then stated that their findings were relevant to decision makers as they provided an understanding of both the direct and indirect impacts of SBW outbreaks on a regional economy.

Mountain pine beetle

The MPB has also received a significant amount of applied economic analysis. An early study from British Columbia by Niemann (Reference Niemann1985) applied CEA to three common control strategies. The aim was to find the strategy that made best use of available funding. He argued that one should start with the cheapest control action and proceed to a more expensive action if any funds were left. He also clarified that, from an economist’s point of view, wood volume at risk was not the main economic concern; it is the volume that could be saved that matters more. Further, he indicated that with more spending on detection and research, additional analysis could be conducted that would reduce the uncertainty of pest management outcomes and improve the decision making by policy makers.

Reed and Errico (Reference Reed and Errico1987) modelled the long-term effects of MPB at both the stand and forest-level in the Quesnel Timber Supply Area in British Columbia. At the stand level, they found that the risk of a pest infestation had very little influence on the land expectation value and the rotation age of lodgepole pine. They attribute this result to the fact that the probability of an infestation is low when stands are immature and because infested timber could be salvaged. At the forest level, they model the stochastic nature of infestations using an average annual infestation rate. While the authors recognised the limitations of this approach they showed it was reasonable over the longer term when infestation intensities were not high.

Miller et al. (Reference Miller, Carlson and Stemeroff1993) presented a CBA with the purpose of determining the benefits and costs of the British Columbia Ministry of Forest’s control programme against the MPB. They found that the programme provided significant net benefits compared to a “no control” approach, but an enhanced programme offered little benefit when compared against the status quo.

Once the most recent MPB epidemic was underway in British Columbia, the attention of research changed from management to predicting the economic impacts that would result in the aftermath of the outbreak. Patriquin et al. (Reference Patriquin, Lantz, Stedman and White2007) developed a computable general equilibrium (CGE) model to investigate the regional economic impacts of MPB in five areas of British Columbia. The model they used was constructed on the basis of both fixed and flexible price approaches. It consisted of six producing sectors and the MPB outbreak was assumed to directly affect timber supply, which then would transfer to other industrial sectors through market activities. The authors demonstrated that all regions were facing a short-term boost in their local economy due to increased salvage. By contrast, in the long run the areas will experience a shortage of timber supply, which will have an associated negative impact on the broader regional economy. However, in a follow up paper, Patriquin et al. (Reference Patriquin, Wellstead and White2008) showed that it may be possible to reduce the negative economic impact of the beetle on a given region by arranging an inter-temporal timber supply trade between regions. Effectively, under this reciprocal wood flow arrangement, regions that were heavily affected by the beetle could allocate their excess salvaged timber to an unaffected region. Correspondingly, the unaffected region would temporarily reduce their annual allowable cut proportionately and return timber to the affected region in the future following salvage.

As the MPB outbreak in the British Columbia Interior progressed, the forest products market was rapidly changing as a result of the financial crisis brought on by the collapse in the United States of America housing market. Schwab et al. (Reference Schwab, Maness, Bull and Roberts2009) modelled the effects of a market downturn on MPB salvage harvesting and the broader forest products industry in British Columbia. They created an agent-based model that consisted of three modules: a resource inventory module, an economic agent module, and a market module. The results projected that timber supply would be significantly reduced resulting in declining stumpage revenues and ultimately a drastically altered structure of the British Columbia forest sector.

Using a dynamic partial equilibrium trade model, Abbott et al. (Reference Abbott, Stennes and van Kooten2009) estimated the impact of MPB-related timber supply shortages on the forest sector. The model incorporated other changes occurring at that time to international forest products markets, namely the introduction of Russian log export taxes, and increases in plantation timber supply in the southern hemisphere and Sweden. From British Columbia’s perspective, the study predicted that the most significant impacts will come from MPB-related timber supply reductions, which will constrain output but will also raise product prices somewhat.

The poor market conditions brought on by the financial crisis created concern that significant pine-beetle affected areas would go unsalvaged. Peter and Bogdanski (Reference Peter and Bogdanski2010) investigated the economic case for salvage harvesting and stand rehabilitation in face of the MPB. They calculated stand-level net present value across a range of site conditions and treatments. The analysis showed that low-value stands with a positive outlook for natural or advance regeneration may generate greater stand value when left unsalvaged. Furthermore, where salvage harvesting was uneconomic, rehabilitation could be justified economically only on sites with high productivity, low treatment costs and a poor outlook for natural regeneration.

At the forest level, Bogle and van Kooten (Reference Bogle and van Kooten2012) demonstrated how a management objective that seeks the best post-salvage inventory state can still achieve salvage objectives, while ensuring timber supplies will be there to sustain the forest sector after the salvage era is over. These same authors (Bogle and van Kooten Reference Bogle and van Kooten2013) demonstrate that the current system of deducting allowances for silviculture costs from stumpage payments has left forest companies with limited incentives to spend on silviculture, particularly when MPB-affected stands have nominal stumpage values that are less than even basic silviculture costs. Framed as a principal-agent problem, Bogle and van Kooten (Reference Bogle and van Kooten2013) demonstrate how greater flexibility could reduce silviculture costs while maintaining forest productivity objectives. The study concludes that new funding mechanisms or incentives should be explored that reward efficiency in achieving silvicultural objectives.

The immense stock of degrading MPB timber also generated significant interest for bioenergy production. Using a regional wood fibre allocation model, Stennes et al. (Reference Stennes, Niquidet and van Kooten2010) showed that expanded bioenergy production in British Columbia would not use standing MPB timber directly but instead would tend to use residuals from the primary forest sector, which had the potential to harm existing users of this fibre (e.g., pulp and pellet mills) by bidding up the price of this supply. Niquidet et al. (Reference Niquidet, Stennes and van Kooten2012) investigated the feasibility of a large wood biomass-fired power plant designed to employ the beetle-killed pine in the Quesnel Forest District, a region at the heart of the MPB outbreak. The focus of their analysis was feedstock costs and showed that average delivered feedstock costs would increase significantly over the life of an electricity power generating facility. The sources of cost inflation were attributed to increasing transportation costs, lower quality wood, and the decreasing availability of cheap roadside residuals in the biomass supply mix over time.

Mountain pine beetle management in British Columbia is now primarily about salvage, whereas Alberta is still focussed on controlling the spread of the beetle. Mathey and Nelson (Reference Mathey and Nelson2010) studied how a resource manager could respond to potential MPB outbreaks using data from west-central Alberta in a 40-year planning exercise. A feature of their model is that it is spatially explicit and it allows for an adaptive response as information changes. The authors termed this modelling approach “planning with recourse”, which involved a repeated two-step process. In the first step a management plan was generated according to a constrained net present value objective function. In the second step, forest dynamics were simulated including a stochastic beetle attack, and the forest plan was re-evaluated. They concluded that the success of a strategy was related to the age-class structure of the forest and the time horizon considered. They also found that increased harvesting levels usually paid off in improved profitability but resulted in a younger forest structure, which likely reduced other forest values such as biodiversity and wildlife habitat. They note that government will likely find it easier to implement strategies involving increased harvest levels in forests with a surplus of older stands, since these strategies also result in improved profits to firms (assuming forest product markets are also favourable).

Other native and non-native pests

One other native Canadian pest that received attention in the economic literature is jack pine budworm, Choristoneura pinus Freeman (Lepidoptera: Tortricidae). Rawat et al. (Reference Rawat, Belli, Smith and Nautiyal1987) attempted to optimise a single pesticide application to control this pest and studied how it may influence the optimal rotation age of jack pine in Ontario. While they noted that quantitative information was lacking for this exercise, they found that immediate pesticide application may not always be the most prudent action once the pest is noticed. Furthermore, they showed that various combinations of pest and forest conditions can drastically alter the economic rotation age.

In terms of non-native pests, Krcmar-Nozic et al. (Reference Krcmar-Nozic, Wilson and Arthur2000) provided a detailed synthesis of research related to the potential impacts of exotic forest pests in North America, which included some economic impacts. However, most of the studies were derived from the United States of America. Aukema et al. (Reference Aukema, Leung, Kovacs, Chivers, Britton and Englin2011) provided an up-to-date summary of the economic impacts of non-native forest insects, focussing on the continental United States of America. Langor et al. (Reference Langor, Cameron, MacQuarrie, McBeath, McClay and Peter2014) describe the potential economic impacts of invasive species on Canada’s boreal zone, noting that even timber impacts are not as straightforward to determine as is sometimes assumed. Nonetheless, little work has been done in Canada.

Yemshanov et al. (Reference Yemshanov, McKenney, de Groot, Haugen, Sidders and Joss2009) is a unique example of a detailed Canadian study that was conducted recently. They assessed the impact of the pine woodwasp (Sirex noctilio), a new alien species detected in the United States of America in 2004 and in eastern Canada in 2005, and evaluated the feasibility of alternative harvesting policies. The analysis was conducted using the Canadian Forest Service Forest Bioeconomic Model (CFS-FBM), in which certain biophysical attributes (i.e., forest growth, timber yields, etc.) and economic analyses (i.e., timber supply and net revenue from forestry activities) were simulated. One of the main contributions of this work was incorporating detailed spatial and temporal dynamics into the analysis. Due to a lack of information about the woodwasp in Canada, an approach that was similar to a model of gypsy moth in the United States of America (Sharov and Liebhold Reference Sharov and Liebhold1998) was used to project the spread of Sirex noctilio. Four disturbance scenarios were examined along with three levels of adaptation: (1) prioritise stands by value as if the infestation did not exist (no adaptation); (2) prioritise stands by value with adjustments as the infestation progresses; and, (3) prioritise stands as in (2) but with an emphasis on harvesting stands closest to infested areas, in an effort to reduce losses. The results indicated that adaptation policies have the potential to significantly reduce economic losses. However, harvesting closer to the infestation was found to be costly in terms of increased transportation costs and harvesting of lower value stands. On the other hand, a policy with no adaptation (1) would likely result in a failure to maintain annual allowable cut levels in the future.

McKenney et al. (Reference McKenney, Pedlar, Yemshanov, Lyons, Campbell and Lawrence2012) quantified the potential economic costs to Canadian municipalities associated with emerald ash borer. Costs included those associated with the removal and replacement of affected trees, pesticide treatments to prevent damage, and overhead costs (e.g., coordination, monitoring, and communication). Scenarios varied according to the rate of spread of the pest, the discount rate and the amount of pesticide applied. Damage was shown to be potentially quite significant, ranging between CAN$265 million to CAN$1177 million. This led the authors to conclude that there is significant economic justification for ongoing management of the emerald ash borer.