Implications of Policy Uncertainty for Innovation in Environmental Technologies: The Case of Public R&D Budgets

  • Margarita Kalamova
  • Nick Johnstone
  • Ivan Haščič
Chapter

Abstract

The role that environmental policy uncertainty can play on innovation in environmental technologies has not been extensively assessed empirically. In this chapter, we seek to assess the impact of environmental policy uncertainty on innovation, using patent data as a proxy for innovation and volatility in public expenditures on ‘environmental’ R&D as a measure of policy uncertainty. Drawing upon a panel data set of 23 OECD countries over the period 1986–2007, support is found for the negative effect of public R&D volatility on innovation. In the base model, a 10% increase in policy uncertainty is seen to cause a 1.2–2.8% decrease in environmental patent activity, whereas a 10% increase in government support for R&D will increase innovation by 2.6–3.9%.

Keywords

Policy uncertainty Innovation Public R&D Patents OECD 

6.1 Introduction

The importance of uncertainty and irreversibility on investment has been well researched in the theoretical literature. However, for some time, these factors seemed to be missing from most empirical research on investment and have created a ‘somewhat disturbing gap between theory and empiricism’ as noted by Pindyck (1991a). While this gap has been filled, some areas remain for which empirical evidence is limited. In particular, the role that environmental policy uncertainty can play on innovation in environmental technologies has not been extensively assessed empirically.

In this chapter, we seek to assess the impact of environmental policy uncertainty on innovation, using patent data as a proxy for innovation and volatility in public expenditures on ‘environmental’ R&D as a measure of policy uncertainty. Indeed, the policy framework in pollution and resource-intensive sectors can be one of the most important factors in the investment decision. For instance, investors in waste-intensive sectors in Europe need to anticipate forthcoming directives which will have an effect on their capital investments. In the context of climate change, the issue may be even more important since government commitments are the outcome of both domestic and international negotiations. In such cases, there can be strong incentives to wait until ‘the policy dust settles’ before adopting a specific investment strategy.

The potential effect of environmental policy uncertainty on incentives to invest in environmental technology arises from the real options literature (see Dixit and Pindyck (1994) for the classic treatment; Pindyck (2007) discusses the specific case of environmental investments). The value of an investment project depends on future output prices, input costs, interest rates, etc. The opportunity cost of the option to invest is a significant component of the firm’s investment decision. The option value increases with the sunk cost of the investment and with the degree of uncertainty over the future price. As Dixit and Pindyck (1994) show in this case, the option to invest will be exercised when output prices exceed the cost of input by an amount equal to the value of keeping the investment option alive. Hence, fluctuations in the value of a project can be traced back to uncertainty in these more basic variables.

The findings of the option-value literature are particularly relevant in the context of investment in innovative activities, such as investment in R&D, because such investments are by nature almost irreversible. Since the costs of these investments cannot usually be recovered if market conditions change, market uncertainty can serve as a significant brake on investment in innovative activities. For instance, in a panel data study of nine OECD countries covering the period 1981–1992, Goel and Ram (2001) find a much sharper adverse effect of uncertainty on R&D investments than on non-R&D (and aggregate) investments. Almost all R&D expenditure on personnel, equipment and materials is irreversible since it is particularly firm-specific or industry-specific, or has the lemons problem as noted by Pindyck (1991b). Thus, irreversibility of investment accentuates the adverse effect of uncertainty (Pindyck 1991b; Dixit and Pindyck 1994).

Importantly, market uncertainty can be compounded by policy uncertainty. For instance, Dixit and Pindyck (1994) show that a case for policy intervention will arise only if firms face a different value of waiting than society as a whole (i.e. if some market failure is associated with the decision process). They study the effect of uncertainty concerning future policy itself and discuss the example of an investment tax credit. They assume that firms will attach more value to waiting because there is a probability that the cost of investment to the firm will fall and find that this policy uncertainty can have a powerful deterrent effect on immediate investment. They conclude that if governments wish to stimulate investment, perhaps the worst thing they can do is to spend a long time discussing the right way to do so. Rodrik (1991), among others, shows that if each year there is some probability that the policy will be reversed, the resulting uncertainty can eliminate any stimulative effect that the policy would otherwise have had on investment. The public economics literature generally confirms the conventional wisdom that tax policy uncertainties can adversely affect firms’ incentives to invest.1

Environmental policy uncertainty has become a significant concern of late. By imposing a price (either explicitly or implicitly) on the costs of pollution emissions, environmental policy is likely to induce innovation as firms seek to meet the policy objectives at least cost. However, if there is uncertainty concerning the stringency, nature and timing of the policy introduced, this can encourage potential innovators to wait before undertaking the necessary investments. It is important to note that environmental policy uncertainty will arise even in an optimal policy setting.

On the one hand, this may be due to uncertainty concerning environmental damages. With unknown damages and increasing information through time, policy conditions may change as the magnitude of the benefits of policy interventions become better known. For instance, there is a large and growing body of literature on the implications of uncertainty regarding climate change damages on the optimal degree of stringency of environmental policy (Baker and Adu-Bonnah 2008). On the other hand, uncertainty with respect to technological conditions may lead to changing environmental policy conditions. Factor substitution possibilities may not be known with any degree of certainty. For instance, in the US Acid Rain Program, the difference between ex ante and ex post estimates of abatement costs has been considerable (Burtraw and Palmer 2004). As such, the optimal level of policy stringency may also change as policymakers acquire information on market responses.

Empirical evidence is mixed concerning the impact of damage and technological uncertainty on the optimal level of investment in R&D on environmental technologies (Baker and Shittu (2006) review much of the recent literature on these two sources of uncertainty in the context of climate policy models). However, the effect of actual policy uncertainty on innovation in environmental technologies has not been examined empirically, although Yang et al. (2008) assess the effects of climate policy uncertainty or fuel choice. In particular, if the future trajectory of this cost is uncertain, option-value theory indicates that individual firms may choose to wait before undertaking investments which seek to identify the means of reducing this cost (i.e. before investing in environmental R&D). Since expectations concerning environmental policy are one of the key determinants of perceived uncertainty over the firm’s planning horizon, policy predictability can play an important role in inducing environmental innovation. In this chapter, we seek to examine formally the proposition that policy uncertainty has slowed investment in environmental innovation. Specifically, drawing upon a database of patent applications from a cross section of 23 OECD countries over the time period 1986–2007, evidence is provided for the negative effect of policy uncertainty of the domestic environmental policy regime on the rate of innovation for environmental technologies. Contrary to previous studies, this chapter makes a novel attempt to measure policy uncertainty by using the coefficient of variation of public R&D expenditures as a proxy for uncertainty.

6.2 Policy Uncertainty and Investment Irreversibility

As noted above, compliance with environmental regulation is often a moving target as environmental regulations are likely to evolve over time. Yet, an investment, once made, has aspects of irreversibility and reversing a regrettable choice is costly (Purvis et al. 1995). Hence, uncertainty over the policy regime can affect incentives to develop and adopt environmental technologies. This policy uncertainty can take different forms:
  • Uncertainty concerning the stringency of the policy and thus the ‘price’ associated with polluting

  • Uncertainty concerning the timing of the introduction of the policy and thus the point at which a ‘cost’ is incurred

  • Uncertainty concerning the nature of the instrument to be used and thus the means by which the cost is incurred

  • Uncertainty concerning the ‘durability’ of the policy and thus the horizon over which the price can be assumed to be in place

There are a small number of studies that address one or more of these aspects of environmental policy uncertainty, indicating that both the rate and direction of innovation can be significantly affected by policy uncertainty.

For instance, there is significant anecdotal evidence in the area of renewable power development to support the hypothesis that uncertainty concerning the time horizon over which investors foresee a given policy to remain in place has played at least as important a role as policy stringency (Söderholm et al. 2007; Wiser and Pickle 1998; Barradale 2008). In particular, Barradale (2008) argues that in the case of the United States, uncertainty concerning annual renewal of the federal production tax credit (PTC) discouraged investment in renewable energy. This finding is supported by anecdotal evidence presented in Wiser and Pickle (1998) concerning both wind and solar power. In a comparison of wind power development in Denmark, Germany and Sweden, Söderholm et al. (2007) argue that the relatively slow pace of development in Sweden is due to instability in the policy framework more than the actual level of support, with a number of different subsidy programmes implemented successively for short periods of time. Interestingly, Barradale (2008) provides evidence that perceived uncertainty is correlated with instrument choice. Investors in the sector believed that renewable energy portfolio standards were more likely to stay in effect long enough to influence long-term investment decisions than depreciation rules, tax credits, feed-in tariffs or production subsidies (which all have direct implications for public budgets).

Isik (2004) analyses the extent to which uncertainty over cost-share subsidy policies aimed at achieving pollution reductions by accelerating the adoption of different farming systems and new technologies in agriculture would impact farmers’ adoption decisions using an option-value model. The author showed that an increase in the probability of an expected public policy leads farmers to delay the adoption of new technologies in order to learn more about market conditions and the value of these technologies. Furthermore, cost-share subsidy policies are found to be more effective when they are immediately offered to farmers and a guarantee provided that they will be removed soon.

Uncertainty concerning the nature of the instrument to be implemented can also have an effect on the rate and nature of innovation. Even if a government is committed to a given environmental objective and provides a credible time frame for its achievement, investors are likely to delay investment until the precise form of the policy instrument is proposed. This is particularly important if the policy options that the government has at its disposal include technology-based standards. In such circumstances, the investor runs the risk of ending up with significant stranded assets if the specific abatement technology is not consistent with permit requirements. The risk is much less if more flexible instruments are introduced. However, even in the case of performance standards, the risk can be considerable if the abatement option adopted does not allow for ex post adjustment of performance levels.

Regulatory uncertainty can also affect the direction of innovation. When choosing between alternative abatement options, the firm must assess the cost of initial capital investment, operating costs and the costs of adjusting production technologies in the face of changing policy conditions. It is important to note that the cost of adjusting production technologies reflects any additional expenditures incurred, minus any salvage value obtained from the resale of capital equipment which is no longer of value to the firm.2 Interestingly, despite regulatory uncertainty, there may be an incentive to invest in end-of-pipe (EOP) abatement rather than more integrated changes in production processes (CPP), even if the latter is a more cost-effective means of mitigating pollution. In the former case (EOP), the abatement decision can be ‘hived off’ from more general production decisions, reducing the probability of being left with stranded assets if there is a change in policy conditions.

The case of coal-fired electricity generation is particularly interesting in this respect. In the face of existing or potential constraints on CO2 emissions, investors in coal-fired electricity generation face a choice between investment in advanced pulverised coal (APC) or integrated gasification and combined cycle plants (IGCC). The capital costs for the former are somewhat lower than for the latter. However, the costs of retrofitting for carbon capture and storage (CCS) are much higher for APC than for IGCC (Bohm et al. 2007). In the presence of regulatory uncertainty over future carbon prices or CCS requirements, there is a value attached to investing in the more ‘flexible’ capital equipment (IGCC). However, uncertainty will slow the delay of the retirement of existing (and more polluting) facilities. Indeed, in a numerical simulation model, Reinelt and Keith (2007) find that under plausible assumptions emissions may be higher when there is regulatory uncertainty than when there is certainty that no regulation will be introduced. The social costs generated can be considerable.

There are a small number of studies that have assessed the role of cost uncertainty arising from either changes in the regulatory regime or volatility inherent to the regulation itself (i.e. permit price volatility or changes in tax levels). For instance, Fuss et al. (2008) also examine the CCS investment decision but focus on the difference between market-driven price uncertainty and policy-driven price uncertainty. The latter is measured as a discrete break in the CO2 price trajectory. They find that market-driven price uncertainty may result in earlier investment in CCS than under conditions of no price uncertainty. However, policy-driven price uncertainty will always delay the CCS investment decision.

Xepapadeas (2001) defines uncertainty as stochastic movements of tradable emissions prices or unpredictable (from the firms’ point of view) policy changes. Moreover, the author accounts for the irreversibility of abatement investment expenses. His analysis yields implications for the regulator concerning the optimal policy design: a regulator can design a policy scheme consisting of two instruments – an emissions tax or tradable permit system and a subsidy on abatement investment. The policy scheme takes uncertainty into account through its dependence on the parameters of the price process and will induce individual firms to undertake the same output and abatement investment under uncertainty that a regulator would have undertaken. Farzin and Kort (2000) theoretically analyse the effects of uncertainty over the size of a tax increase at a certain future date and uncertainty over the timing of a known tax increase. Their results suggest that though both types of uncertainties affect the optimal abatement investment path, the effect of the former may be more pronounced, especially when investment is irreversible. Interestingly, they show that a credible threat of accelerating the tax increase can further boost the firm’s abatement investment.

In one of the few formal empirical studies, Löfgren et al. (2008) assess Swedish firms’ investments in pollution abatement technology related to SO2 emissions. In their model, the price of the polluting fuel is the major source of uncertainty facing the firm, drawing upon a panel of firms from the Swedish pulp and paper industry and the energy and heating sector, and their sulphur dioxide emissions over the period 2000–2003. The results indicate that in the presence of uncertainty over the price of the polluting fuel, the hurdle rates – i.e. the multiplier of the price of the polluting fuel relative to a condition of perfect information necessary to trigger investment in the less pollution technology – are between 2.7 and 3.1 for the pulp and paper sector and 3.4 and 3.6 in the energy and heating industry. Interestingly, they note that there are differences between firms that invest in EOP vs. CPP technologies, but no firm conclusions are drawn regarding the role of uncertainty in guiding the decision.

In a study on the US pulp and paper sector, Maynard and Shortle (2001) assessed the effect of protracted uncertainty concerning the development of the US EPA’s Cluster Rule (which targeted dioxins) on adoption of less polluting technologies. They examined three abatement options: extended delignification (ED), oxygen delignification (OD) and more advanced elemental chlorine-free bleaching (ECF). The different options have interesting characteristics. While ED and OD are more integrated in the production process than ECF, which can be considered a form of end-of-pipe technology, the cost of implementing ECF is less if the plant already has invested in ED or OD. Using a double-hurdle model, they find that the uncertainty surrounding the policy encouraged investors to ‘wait and see’ before undertaking the investments in both extended or oxygen delignification or elemental chlorine-free bleaching. Prior investment in ED or ED affected the decision to invest in ECF.

Theoretical arguments and empirical evidence indicate that the effects of frequent and unpredictable policy changes on long-term investments can, therefore, be considerable.

6.3 Hypothesis

In the environmental context, unlike many other areas, the viability of a specific investment is dependent upon a specific policy regime remaining in place. This is a major risk that inventors will need to evaluate before deciding on the investment project. Companies must absorb significant risk during the research and development phase of a product if there is some uncertainty that a particular policy will apply to their project when it becomes commercially viable. Even where policies survive, attempts at legislative intervention, agency and/or court rulings can significantly alter a policy’s applicability and implementation. Since unpredictability of these policies provides some uncertainty to the profitability of innovation efforts, companies will be reluctant to innovate. The empirical hypothesis can therefore be stated as follows: Uncertainty over environmental policy will have a negative impact on a firm’s decision to innovate in environmental technologies.

Our measure of innovation is counts of patent applications for environmental technologies, discussed in further detail below. As a measure of policy uncertainty, we use volatility in public expenditures on environmental R&D. This includes both direct government expenditures for R&D undertaken in government and publicly funded university laboratories as well as the provision of financial support (grants, tax credits, etc.) for R&D undertaken by the private sector and other organisations.

This measure should reflect at least two aspects of uncertainty. First, since some form of public fiscal support is usually necessary for privately undertaken R&D projects in environmental technologies to be feasible at all, this measure will reflect variation in the cost of the investment. Second, since public R&D targeted at a specific field can be considered as a signal of related public policy objectives, volatility in such expenditures can be used as a measure of commitment.

6.4 Data and Empirical Analysis

In this study, patent data are used to construct a proxy measure of environmental innovation. Patent data have been used as a measure of technological innovation because they focus on outputs of the inventive process (Griliches 1990; OECD 2009). This is in contrast to many other potential candidates (e.g. research and development expenditures, number of scientific personnel) which are at best imperfect indicators of the innovative performance of an economy since they focus on inputs. Moreover, patent data provide a wealth of information on the nature of the invention and the applicant; the data is readily available and discrete (and thus easily subject to statistical analysis). Significantly, there are very few examples of economically significant inventions which have not been patented (Dernis et al. 2001).

The data used to construct this indicator were taken from the OECD Patents Statistics database3 based on counts of patent applications in key areas of environmental technology – air pollution abatement, water pollution abatement, solid waste management, soil remediation and environmental monitoring technologies. (See Appendix for a list of IPC classes used to identify the relevant patented inventions.) The dependent variable represents the number of patent applications deposited at the European Patent Office, classified by inventor country4 and priority year.5 To test the empirical hypothesis, the following model is estimated:
$$ { \text{ENVPA}}{{\text{T}}_{i,t}} = f({\text{TOTAL}}{\text{PATENT}}{{\text{S}}_{i,t}},{\text{GBAOR}}{{\text{D}}_{{\rm{EN}}{{\rm{V}}_{i,t}}}},{\text{POLICY}}{\text{UNCERTAINT}}{{\text{Y}}_{i,t}}) + {\varepsilon_{i,t}} $$
(6.1)
where i indexes country and t stands for year. The dependent variable is measured by the number of patent applications in environmental technology as described above.

It is important to control statistically for differences in the propensity to innovate and patent across countries. In order to capture the effect of such factors (which are not specific to environmental technologies), we include the variable TOTAL PATENTS reflecting the total number of patent applications deposited at the EPO filed across the whole spectrum of technological fields (not only environmental). This variable thus controls for differences in a country’s general research capacity as well as changes in general propensity to patent over time and across countries. Ideally, we would estimate the model using a two-stage procedure where total patenting activity is first estimated. This approach was followed (Johnstone et al. 2012) and results from the two-stage estimation were seen to be closely comparable with those from a reduced-form model. Since many observations would be lost with such an approach, in this case, we have decided to adopt this strategy.6 The sign on this variable is expected to be positive.

In previous work on the determinants of environmental innovation, relative policy stringency has been included as the principal environmental policy factor (Brunnermeier and Cohen 2003; Lanjouw and Mody 1996; Johnstone et al. 2010). The relative stringency of environmental policy is thought to induce innovation by changing relative factor prices or introducing production constraints (Hicks 1932). However, measurement of this effect is complicated because cross-country (or cross-sectoral) data on regulatory stringency are rarely available or are not commensurable. Moreover, public policies typically target specific environmental impacts (pollutants) using a specific policy instrument. This chapter deals with a broadly defined (environmental) technology and hence covers multiple impacts and potentially a wide spectrum of policy instruments and sectors. Moreover, it operates in a cross-country context. Previous studies have used data on pollution abatement and control expenditures of the private sector (PACE) as well as on perceived stringency (survey by the World Economic Forum) to measure the stringency of environmental policy regimes. However, the first variable consists of large numbers of missing observations, whereas the second one is available for a very short period of time only (2001–2007).

As noted, in this study, we use government budget appropriations and outlays for R&D (GBAORD). GBAORD data is disaggregated by socio-economic objective, including control and care of the environment7 (GBAORDENV). This is applied as a proxy of policy stringency. The data are taken from the OECD Research and Development Statistics database. The sign of this variable is expected to be positive. More specifically, it covers total government appropriations or outlays for R&D (GBAORD), reflecting not only government-financed R&D performed in government establishments but also government-financed R&D in the other three national sectors (business enterprise, private non-profit, higher education) as well as abroad (including international organisations) (OECD 2002).

In this chapter, the key explanatory variable is a measure of environmental policy uncertainty (POLICY UNCERTAINTY). While there are no studies on the uncertainty of environmental policy, there are several papers examining the effect of market volatility on investment in general. Typical measures of uncertainty in these studies are n-year moving standard deviation or moving average deviation of the variable of interest (e.g. of inflation in Goel and Ram (2001)) or its variance (of firm’s daily stock returns, in Bloom and Van Reenen (2002)). Although the World Economic Forum survey asked managers for their perceptions of stability in environmental policy, the index of uncertainty is available only for the time period 2001–2007. This chapter makes a novel attempt at measuring policy uncertainty and estimating its effect on environmental innovation by using the coefficient of variation of GBAORDENV as a proxy of uncertainty. Following the method of Czarnitzki and Toole (2011) for calculating market volatility (uncertainty), the coefficient of variation is calculated for each country across time based on a pre-sample data of 5 years (over the time period 1981–1985):
$$ {{\text{POLIC}}{{\text{Y}}\;{{\rm{UNCERTAINT}}{{\rm{Y}}_{i,t}}}} = \frac{{{{\sqrt {{\frac{1}{5}\sum_{s = 0}^4\left[ {{\text{GBAOR}}{{\text{D}}_{{\rm{EN}}{{\rm{V}}_{i,t - s}}}} - \left( {\frac{1}{5}\sum_{s = 0}^4{\text{GBAOR}}{{\text{D}}_{{\rm{EN}}{{\rm{V}}_{i,t - s}}}}} \right)} \right]}} }^2}}}{{\frac{1}{5}\sum_{s = 0}^4{\text{GBAOR}}{{\text{D}}_{{\rm{EN}}{{\rm{V}}_{i,t - s}}}}}}} $$
(6.2)
The theoretical minimum value is 0, and the maximum is 1. The correlation between the levels and volatility of R&D is approximately −0.4. For the regression analysis, we use the lag of the policy uncertainty variable. Based on our principal hypothesis, this variable is expected to have a negative sign. Figure 6.1 presents the relationship between the level and volatility of R&D spending. It is interesting to note that volatility is somewhat higher for countries that spend a relatively lower % of GBAORD on the environment.
Fig. 6.1

Relationship between the level and volatility of ENV R&D spending

All the residual variation in the dependent variable is captured by the error term εit. A negative binomial model is used to estimate the model (for details on count data models, see, e.g. Cameron and Trivedi 1998; Maddala 1983; Hausman et al. 1984). Descriptive statistics for the estimation sample of 23 OECD countries over the period 1986–2007 are provided in Table 6.1.
Table 6.1

Descriptive statistics

Variable

Number of observations

Mean

Std. dev.

Min

Max

Environmental patents

422

78.06833

130.8575

0

537.331

TOTAL PATENTS (in ‘000s of patents)

422

3.954042

7.114198

0.010417

34.48414

GBAORDENV (in millions of 2000 USD)

422

130.1552

169.6466

0.272

672.48

POLICY UNCERTAINTY (lag)

422

0.196477

0.141237

0.009715

0.809435

Table 6.2 presents our regression results. In a first step, we consider the effect of policy uncertainty in a pooled estimation (see column 1). Then, in column 2, we include country fixed effects. Column 3 considers the lag of GBAORDENV instead of its contemporaneous value. Heteroskedasticity-robust standard errors are reported in parentheses. To summarise, the baseline results provide strong and consistent support for the hypothesis according to which a higher policy uncertainty will discourage innovation in environmental technologies. In Table 6.2, POLICY UNCERTAINTY is significant and negative across all specifications, whereas the GBAORD variable is significant and positive. Furthermore, there is little difference between the regressions using lagged GBAORDENV or its contemporaneous value (columns 2 and 3, respectively). The coefficient on the TOTAL PATENTS variable is positive and highly significant suggesting that patenting activity in the selected ‘environmental’ technologies is also explained by variation across countries and over time in patenting activity overall. In order to compare the relative magnitude of the coefficients, the elasticities have been calculated and are presented in Fig. 6.2. Thus, a 10% increase in policy uncertainty will cause a 1.2–2.8% decrease in environmental patent activity in the models with fixed effects (columns 2 and 3) and without fixed effects (column 1), respectively. At the same time, a 10% increase in government support for R&D will increase innovation by 2.6–3.9%.
Table 6.2

Baseline results of the effect of policy uncertainty on innovation

 

(1)

(2)

(3)

TOTAL PATENTS

0.1085221***

0.0157791***

0.012334***

(0.0102444)

(0.0036798)

(0.0037922)

GBAORDENV

0.0030052***

0.0020716***

(0.0002866)

(0.0002344)

GBAORDENV (lag)

0.0020853***

(0.0002536)

POLICY UNCERTAINTY

−1.44427***

−0.6286821***

−0.6383595***

(0.3385673)

(0.1577657)

(0.1575273)

Intercept

2.954181***

4.407299***

4.506464***

(0.0834035)

(0.1291679)

(0.1318719)

Country fixed effects

No

Yes

Yes

Number of obs.

422

422

422

Log pseudolikelihood

−1,899.8567

−1,447.43

−1,449.6019

(Prob > Chi2)

0

0

0

Note: (1) *** – significant at 1% level, ** – significant at 5% level, * – significant at 10% level; (2) Dependent variable is ENVPAT

Fig. 6.2

Elasticities for the level and volatility of R&D spending

Since the main explanatory variable in this study, POLICY UNCERTAINTY, is based on a government budget indicator (GBAORD), it is important to account for structural reforms occurring in the OECD economies which might have affected the reporting of public data. If we fail to account for potential breaks in the GBAORD series, the volatility of the uncertainty variable will be inflated. Indeed, we can identify several breaks in the GBAORD series, including the 1991 German reunification as well as the 1992 French economic reform to boost consumption and the record peak in surplus of the US federal budget in 2000. Therefore, in a subsequent step, we run the estimation on a reduced sample that accounts for structural breaks in the GBAORDENV variable and thus in the uncertainty measure. In general, the signs and significance levels of the explanatory variables remain the same as in the baseline estimation (except for the uncertainty measure in column 3), but the magnitudes of their estimated coefficients and elasticities change slightly. Thus, a 10% increase in policy uncertainty will cause a 0.6% (in the models with FEs) to 5.3% (without FEs) decrease in innovation activity whereas a 10% increase in GBAORDENV will increase innovation by 1.2–3.3%. Further robustness checks include the use of the logarithm of government R&D expenditures as well as their share in the country’s gross domestic product as explanatory variables. The regression analysis delivers the same qualitative results as in the baseline estimation (Table 6.3).
Table 6.3

Tests of robustness

 

(1)

(2)

(3)

TOTAL PATENTS

0.1281067***

0.0337766***

0.034552***

(0.0090737)

(0.0073192)

(0.0078704)

GBAORDENV

0.0026519***

0.001165***

(0.0004066)

(0.0003138)

GBAORDENV (lag)

0.0010146***

(0.0003031)

POLICY UNCERTAINTY

−3.007049***

−0.3609491*

−0.3143938

(0.5914218)

(0.2010824)

(0.1997188)

Intercept

3.077275***

4.53315***

4.606643***

(0.1097179)

(0.1347338)

(0.1364057)

Country fixed effects

No

Yes

Yes

Number of obs.

295

295

295

Log pseudolikelihood

−1,298.3554

−923.68148

−926.60404

(Prob > Chi2)

0

0

0

Note: (1) *** – significant at 1% level, ** – significant at 5% level, * – significant at 10% level; (2) Dependent variable is ENVPAT

6.5 Conclusions and Policy Implications

Uncertainty associated with a country’s environmental policy – whether in terms of stringency, timing, nature or durability – will result in less innovation in environmental technologies. It may also bend the direction of innovation in a suboptimal manner. Since the planning horizon for investments in innovation is particularly long and the risk of being left with stranded assets is great, such investment decisions are likely to be significantly affected by policy uncertainty. The consequences can be manifold and include reduced and distorted investment in environmental R&D, delayed retirement of older facilities, suboptimal technology adoption choices and increased emission rates. Conversely, the more predictable a policy regime is, the more likely innovation is to take place and the more likely that this innovation will be directed in an optimal manner. This implies that governments should behave in a predictable manner if they wish to induce innovations that achieve environmental objectives at lower cost. Frequently changing policy conditions come at a cost. Uncertain signals give investors strong incentives to postpone investments, including the risky investments which lead to innovation. There is an advantage to ‘waiting’ until the policy dust settles. As such, by adding to the risk that investors face in the market, policy uncertainty can serve as a ‘brake’ on innovation, both in terms of technology invention and adoption. It is important to note that changing the policy parameters does not necessarily provide more uncertainty to investors as long as this is done in a predictable manner (e.g. periodic adjustments made in response to market developments). This implies that governments have an interest to behave in a predictable manner if they wish to induce innovations that achieve environmental objectives at lower cost. Moreover, these effects of policy uncertainty can be long-lived, stretching well beyond the period of uncertainty, since future investment decisions will be affected by the credibility of signals given by policymakers in the past. A history of abrupt policy changes can discourage future investment long after the period of uncertainty has passed. Credibility is hard-won. However, it should be recognised that in some cases policy uncertainty can arise from the acquisition of new information. Damages may be higher or lower than initially foreseen, encouraging the use of more or less stringent policies. Similarly, abatement costs may be higher or lower than initially foreseen. In such cases, there is a trade-off between changing environmental objectives to reflect the new information and keeping incentives constant in order to reduce uncertainty. One possibility for mitigating the impacts of such trade-offs on the predictability of the policy signal is to design environmental policy in a manner that such ‘exogenous’ sources of uncertainty are explicitly built in alongside other policy parameters.

Footnotes

  1. 1.

    Studies exploring this effect include Rodrik (1991) and Aizenman and Marion (1993), among others.

  2. 2.

    Note that the resale value may be zero if the capital is specific, and the regulatory change impacts on all potential adopters.

  3. 3.
  4. 4.

    ‘Fractional’ counts are generated in cases when inventors from multiple countries are listed.

  5. 5.

    ‘Priority date’ indicates the earliest application date worldwide (within a given patent family).

  6. 6.

    In the sample used for econometric analysis, storage patents represent on average only 0.2% of total patents. Nevertheless, in order to avoid any concern over possible endogeneity, regressions are estimated considering the difference between the patent total and the dependent variable.

  7. 7.

    This covers research into the control of pollution, aimed at the identification and analysis of the sources of pollution and their causes and all pollutants, including their dispersal in the environment and the effects on man, species (fauna, flora, microorganisms) and the biosphere. Development of monitoring facilities for the measurement of all kinds of pollution is included. The same is valid for the elimination of all forms of pollution in all types of environment.

References

  1. Aizenman, J., & Marion, N. P. (1993). Policy uncertainty, persistence and growth. Review of International Economics, 1(2), 145–163.CrossRefGoogle Scholar
  2. Baker, E., & Adu-Bonnah, K. (2008). Investment in risky R&D programs in the face of climate uncertainty. Energy Economics, 30, 465–486.CrossRefGoogle Scholar
  3. Baker, E., & Shittu, E. (2006). Profit-maximizing R&D in response to a random carbon tax. Resource and Energy Economics, 28, 160–180.CrossRefGoogle Scholar
  4. Barradale, M. J. (2008). Impact of policy uncertainty on renewable energy investment: Wind power and PTC (Working Paper 08-003). US Association for Energy Economics.Google Scholar
  5. Bloom, N., & Van Reenen, J. (2002). Patents, real options and firm performance. The Economic Journal, 112, C97–C116.CrossRefGoogle Scholar
  6. Bohm, M. C., Herzog, H. J., Parsons, J. E., & Sekar, R. C. (2007). Capture-ready coal plants: Options, technologies and economics. International Journal of Greenhouse Gas Control, 1, 113–120.CrossRefGoogle Scholar
  7. Brunnermeier, S. B., & Cohen, M. A. (2003). Determinants of environmental innovation in US manufacturing industries. Journal of Environmental Economics and Management, 45, 278–293.CrossRefGoogle Scholar
  8. Burtraw, D., & Palmer, K. (2004). SO2 Cap-and-trade program in the United States. In W. Harrington, R. Morgenstern, & T. Sterner (Eds.), Choosing environmental policy: Comparing instruments and outcomes in the United States and Europe. Washington, DC: RFF Press.Google Scholar
  9. Cameron, A. C., & Trivedi, P. K. (1998). Regression analysis of count data. Cambridge: Cambridge University Press.Google Scholar
  10. Czarnitzki, D., & Toole, A. A. (2011). Patent protection, market uncertainty, and R&D investment. The Review of Economics and Statistics, 93(1), 147–159.CrossRefGoogle Scholar
  11. Dernis, H., Guellec, D., & van Pottelsberghe de la Potterie, B. (2001). Using patent counts for cross-country comparisons of technology output. STI mimeo, Paris: Organisation for Economic Co-operation and Development. http://www.oecd.org/dataoecd/26/11/21682515.pdf
  12. Dixit, A. K., & Pindyck, R. S. (1994). Investment under uncertainty. Princeton: Princeton University Press.Google Scholar
  13. Farzin, Y. H., & Kort, P. M. (2000). Pollution abatement investment when environmental regulation is uncertain. Journal of Public Economic Theory., 2(2), 183–212.CrossRefGoogle Scholar
  14. Fuss, S., Szolgayova, J., Obersteiner, M., & Gusti, M. (2008). Investment under market and climate policy uncertainty. Applied Energy, 85, 708–721.CrossRefGoogle Scholar
  15. Goel, R. K., & Ram, R. (2001). Irreversibility of R&D investment and the adverse effect of uncertainty: Evidence from the OECD countries. Economics Letters, 71(2), 287–291.CrossRefGoogle Scholar
  16. Griliches, Z. (1990). Patent statistics as economic indicators: A survey. Journal of Economic Literature, 28(4), 1661–1707.Google Scholar
  17. Hausman, J., Hall, B. H., & Griliches, Z. (1984). Econometric models for count data with an application to the patents-R&D relationship. Econometrica, 52, 909–938.CrossRefGoogle Scholar
  18. Hicks, J. R. (1932). The theory of wages. London: Macmillan.Google Scholar
  19. Isik, M. (2004). Incentives for technology adoption under environmental policy uncertainty: Implications for green payment programs. Environmental and Resource Economics, 27, 247–263.CrossRefGoogle Scholar
  20. Johnstone, N., Hascic, I., & Kalamova, M. (2010). Environmental policy characteristics and technological innovation. Economia Politica, 2, 277–301.Google Scholar
  21. Johnstone, N., Haščič, I., Poirier, J., Hemar, M., & Michel, C. (2012). Environmental Policy Stringency and Technological Innovation: Evidence from Survey Data and Patent Counts. Applied Economics, 44, 2157–2170.Google Scholar
  22. Lanjouw, J. O., & Mody, A. (1996). Innovation and the international diffusion of environmentally responsive technology. Research Policy, 25(5), 49–571.Google Scholar
  23. Löfgren, Å., Millock, K., & Nauges, C. (2008). The effect of uncertainty on pollution abatement investments: Measuring hurdle rates for Swedish industry. Resource and Energy Economics, 30, 475–491.CrossRefGoogle Scholar
  24. Maddala, G. S. (1983). Limited-dependent and qualitative variables in econometrics. Cambridge: Cambridge University Press.Google Scholar
  25. Maynard, L. J., & Shortle, J. S. (2001). Determinants of cleaner technology investments in the U.S. bleached kraft pulp industry. Land Economics, 77(4), 561–576.CrossRefGoogle Scholar
  26. OECD. (2002). Frascati manual 2002: Proposed standard practice for surveys on research and experimental development. Paris: OECD.Google Scholar
  27. OECD. (2009). OECD patent statistics manual. Paris: OECD.Google Scholar
  28. Pindyck, R. S. (1991a). Irreversibility and the explanation of investment behavior. In D. Lund & B. Oksendal (Eds.), Stochastic models and option values. New York/Amsterdam: Elsevier Science Publishers/North-Holland.Google Scholar
  29. Pindyck, R. S. (1991b). Irreversibility, uncertainty, and investment. Journal of Economic Literature, 29, 1110–1152.Google Scholar
  30. Pindyck, R. S. (2007). Uncertainty in environmental economics. Review of Environmental Economics and Policy, 1, 45–65.CrossRefGoogle Scholar
  31. Purvis, A., Boggess, W. G., Moss, C. B., & Holt, J. (1995). Technology adoption decisions under irreversibility and uncertainty: An ‘ex ante’ approach. American Journal of Agricultural Economics, 77(3), 541–551.CrossRefGoogle Scholar
  32. Reinelt, P., & Keith, D. (2007). Carbon capture retrofits and the cost of regulatory uncertainty. Energy Journal, 28(4), 101–128.Google Scholar
  33. Rodrik, D. (1991). Policy uncertainty and private investment in developing countries. Journal of Development Economics, 36, 229–242.CrossRefGoogle Scholar
  34. Söderholm, P., Ek, K., & Pettersson, M. (2007). Wind power development in Sweden: Global policies and local obstacles. Renewable and Sustainable Energy Reviews, 11, 365–400.CrossRefGoogle Scholar
  35. Wiser, R. H., & Pickle, S. J. (1998). Financing investments in renewable energy: The impacts of policy design. Renewable and Sustainable Energy Reviews, 2, 361–386.CrossRefGoogle Scholar
  36. Xepapadeas, A. (2001). Environmental policy and firm behavior: Abatement investment and location decisions under uncertainty and irreversibility. In C. Carraro & G. E. Metcalfe (Eds.), Behavioral and distributional effects of environmental policy (pp. 281–308). Chicago: University of Chicago Press.Google Scholar
  37. Yang, M., Blyth, W., Bradley, R., Bunn, D., Clarke, C., & Wilson, T. (2008). Evaluating the power investment options with uncertainty in climate policy. Energy Economics, 30, 1933–1950.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2012

Authors and Affiliations

  • Margarita Kalamova
    • 1
  • Nick Johnstone
    • 1
    • 2
  • Ivan Haščič
    • 1
  1. 1.Organisation for Economic Co-operation and DevelopmentOECD Environment DirectorateParis Cedex 16France
  2. 2.Empirical Policy Analysis UnitOECD Environment DirectorateParis Cedex 16France

Personalised recommendations