Get Complete Project Material File(s) Now! »
Uncertainty, irreversibility and high discount rates
By contrast, Hassett and Metcalf (1993) argue that the so-called energy paradox is in reality a optimal response to first, uncertainty about the return on investment (i.e. the energy savings following the adoption of equipment), and second, the irreversibility of the investment (as insulation for example). Indeed, the uncertainty on the evolution of energy prices leads to an uncertainty on energy-savings that will be realized following the investment. Agents have to forecast future energy prices to appraise the profitability of the investment. Moreover, once the investment is undertaken, it cannot be sold if the energy prices fall and the investment becomes unprofitable. Therefore, it is prudent for an agent to wait to get information about energy price trends.
Given uncertainty and irreversibility, agents use high implicit discount rates for energy-saving investment, i.e. the present value of future energy savings is low. A literature review that empirically estimates the implicit discount rates used for energy saving investment show that they substantially exceed the maximum discount rate that consumer would be expected to apply (using the rate of return available on investments of similar risk) (Sanstad et al., 1995). Hassett and Metcalf (1993) find that the discount rate used for energy saving investment exceeds the conventional estimate by a factor of four. Therefore, agents require a largely higher return on investment for energy saving equipment than for other kinds of investment to undertake the project.
In terms of policy implication, this finding stresses the fact that providing more information about benefits of energy saving investments is not sufficient to induce investment. Hasset and Metcalf (1993) show that increased benefits of an investment are also likely to only have a 8 small effect. They simulate the impact of 15% tax credit for energy saving investments and show that the effect of such a policy is dramatically attenuated because of uncertainty. Given these observations, it seems that standards could be efficient to decrease energy consumption because they have no link with the agents’ perception of future energy savings following an investment. Thermal regulations for new constructions and then for renovations have been introduced in France for the first time respectively in 1974 and 2008.
Public policy efficiency vs. green paradox, free riding and rebound effect
The energy paradox leads to under-investment in energy saving equipment. Public policy intervention is then necessary to induce renovations. Public policy can lower some of these barriers and help agents to undertake energy-saving investments, in order to finally significantly reduce energy consumption and greenhouse gas emissions. In recent years, several measures have been implemented in France, (1) informative measures, with for example the presence of eco-label on appliances or bulb-light to inform consumers on the energy efficiency level of the equipment, or information campaigns to rise households sensitivity about energy-savings; (2) financial measures to induce households to adopt renewable energies or improve the housing quality; it can be zero rate bank loans, subsidies or tax credits, or reduced rate VAT; (3) and regulatory measures, such as the thermal regulation on new constructions or renovations, labels, or the requirement to indicate the energy quality of housing when it is sold or offered for rent.
One of the most famous measures is the tax credit. It allows part of the expenses of energy saving renovations to be deducted from income taxes. From 2005 to 2008, 4.2 million French households received a tax credit (Clerc and Mauroux, 2010) and this represents a significant cost: public cost reached €7.8 billion during this period and €4.2 billion during 2009–2010. However, several behaviors can undermine the effect of environmental policies.
Free riding and spillover effects
Financial measures have to be implemented carefully, because of potential free riding. Free riders are households that obtain for example a subsidy to undertake a renovation that would have made even in the absence of public policy. Recent literature estimates the extent of the free-riding effect from 50% to 92%. Grösche and Vance (2009) use a cross-section of data from the 2005 German Residential Energy Consumption Survey to evaluate this effect. They define free riding as a situation in which a household’s willingness to pay for renovations exceeds its cost under no policy action, and show that such a free riding occurs in 50% of the cases. In an original study, Grösche et al. (2009) simulate the effect of grants on renovation choices using revealed preference data on home renovations from Germany’s residential sector. They find that if every eligible household had behaved rationally and applied for the grant, 92% of the program expenses would have been awarded to free riders. Malm (1996) also finds an important free riding effect. He investigates the impact of subsidies on the purchase of high-efficiency heating systems and estimates it at 89%.
However, some spillover effects can reduce free riding (Eto et al., 1995; Rosenow and Galvin, 2013). Such effects correspond to additional products being installed, as a result of the program but not through the program. Few studies focus on this point, but a recent evaluation shows that spillover effects can be substantial (NYSERDA, 2012).
The policies already presented have the objective to induce investment. However, the adoption of energy saving technologies is not necessarily followed by a reduction of energy consumption. It appears that investment in a new technology such as insulation improvement can entail a change in household behavior (e.g., increase in the temperature target), which at least partially offsets the beneficial effects of the technology. This is called the direct rebound effect (for a review see International Risk Governance Council, 2013). One explanation is that people tend to consume more energy services when it is less expensive. Therefore, the rebound effect reduces or offset the impact of environmental measures on energy consumption and greenhouse gas emissions. In a large survey, Greening et al. (2000) find that a 100% increase in energy efficiency led to an estimated rebound of 0%–50% for residential end uses. Also, Alberini et al. (2013b) examine household energy consumption in Maryland and show that the larger the subsidy obtained for the adoption of energy saving equipment, the less the electricity reduction, and this result may be explained by the rebound effect.
Moreover the rebound effect can have indirect impact (Schipper and Grubb, 2000). When energy services are less costly, households have more income and can increase the demand for other goods that require energy for production or use. Druckman et al. (2010) simulate the effect of a set of abatement actions of carbon emissions in UK, using different scenarios. On average, the indirect rebound effect is estimated at 34% of the anticipated GHG emissions reductions (this means that only two thirds of the anticipated GHG emissions reductions are likely to be achieved). These authors also show that in the best case, it may be only 12% but that in extreme cases backfire may occur. Backfire means that carbon emissions increase, instead of decreasing.
An economy-wide rebound effect may also exist and take into account a wide range of effects at the macroeconomic level. Gillingham et al. (2013) explain this effect at a worldwide level using the example of fuel standards on vehicles in the United States, which can lead to a decrease in world oil prices causing in turn an increase in oil demand in other countries. The estimates of this economy-wide rebound effect varies considerably across countries, depending on the model used (computational general equilibrium model, macroeconomic model) and on the variables considered. However, results are generally greater than 37%, with most studies finding larger rebounds or backfire (Sorrell, 2007; International Risk Governance Council, 2013). For example, Barker et al. (2007) examine the rebound effect in the UK related to energy efficiency policies between 2000 and 2010 and show that it was not large enough to prevent a significant decrease in energy consumption and greenhouse gas emissions. They estimate simultaneously the indirect and economy-wide effects using a macroeconomic model and they obtain that the rebound is around 11% on average across all sectors of the economy, i.e. the reduction in energy demand is 11% less than expected. The direct rebound effect is around 15% leading to a total rebound of 26% of the expected reduction of energy demand.
Some environmental policies could become inefficient because of the existence of a green paradox (Sinn, 2008). Instead of decreasing greenhouse gas emissions, measures that aim at decreasing fossil energy demand (as tax carbon or subsidies on renewable energies) could increase pollution and accelerate climate change at least in the short run. Using a Hoteling model, Sinn (2008) shows that the introduction of a carbon tax that rises over time (note that green paradox never happens for the optimal tax path) can indeed have a negative effect. Since the tax will increase the price of the fossil energy over time, producers have incentive to extract and sell the resource immediately. Such a policy therefore accelerates environmental damages. Van der Ploeg and Withagen (2010) show that the green paradox occurs for relatively expensive but clean energy (such as solar or wind). If the government introduces subsidies on solar or wind energy, this leads to an overconsumption of oil and gas, i.e. to a more rapid depletion of these energies. In this case, the energy paradox is confirmed and it has negative effect on climate change. In contrast, there is no evidence of the energy paradox if the clean energy is sufficiently cheap relative to marginal global warming damages (as nuclear energy). In this case, it is attractive to leave fossil fuels unexploited and thus limit CO2 emissions. Grafton et al (2010) find that biofuel subsidies could lead to green paradox depending on several variables, such as the demand and supply elasticities, the expected change in the measure, the technological change in extraction and extraction cost. It seems that the green paradox can also occur when a climate policy is announced in advance and the implementation date is uncertain. Indeed, between these two dates (that of the announcement and that of the implementation) the use of fossil energy and thus the greenhouse gas emissions increase (Smulders et al. 2010).
Given all these effects that can undermine the effectiveness of environmental policies, we can wonder whether French environmental policies are sufficiently efficient to reach the ambitious objectives set by Grenelle Act. This doctoral research aims at providing insights on these issues. We evaluate these measures first at a national level using a simulation model (chapter 2) and second, we focus on one measure and observe its impact on households’ behavior using an econometric approach (chapter 3).
In chapter 2, we test the impact of some existing policies (tax credit, zero rate bank loans, subsidies, and VAT) and of one potential policy (bonuses). We combine several approaches found in the literature and model energy consumption dynamics resulting from both the housing stock dynamics (including three end-uses: heating and hot water, lighting, and appliances) and the energy saving investment decisions. This study produces three major outputs: (1) an estimation of French residential energy consumptions and of GHG emissions until 2050, (2) an assessment of the impact of environmental policies compared to the public cost, and (3) proposals of different means to reach the objectives set out in the Grenelle Act. Results show that current policies are effective in the sense that they have enabled a decrease in energy consumption and in GHG emissions over recent years. A tax credit seems to be one of the most effective policy measures. However, existing policies alone will not ensure that the objectives set for 2050 will be reached. Additional public expenditures are necessary to achieve these goals.
Table of contents :
Chapter 1. What matters in Residential Energy Consumption? Evidence from France
1.4.1. Methodology issues
1.4.2. Analysis method: discrete and continuous choices model
1.7.2. Heating system choice estimations
1.7.3. Enercy consumption estimations
Chapter 2. Evaluation of the impact of environmental public policy measures on energy consumption and greenhouse gas emissions in the French residential sector
2.2. Public policy
2.3.1. Modeling the dynamics of the housing stock
2.3.2. Modeling energy consumption and GHG emissions
18.104.22.168. Energy consumption and GHG emissions related to heating and hot water
22.214.171.124. Energy consumption related to appliances
126.96.36.199. Energy consumption related to lighting
2.4.1. The effects of current environmental policies
188.8.131.52. Quantitative results of basic variables in the reference scenario
184.108.40.206. The effects of current public policies
2.4.2. Comparison of policies
2.4.3. Public policy measures to achieve the Grenelle I goals
2.5. Sensitivity analysis
2.7.1. Introduction of a carbon tax
2.7.2. Change in the evolution of the population
2.7.3. Variables and data
2.7.4. Sensitivity analysis
Chapter 3. Environmental fiscal incentives: Effectiveness or free-riding effect? An econometric evaluation of the French energy tax credit
3.3.1. An evaluation problem
3.3.2. Estimation strategy
220.127.116.11. Renovation rate
18.104.22.168. Renovation expenditures
3.4.1. Propensity score
3.4.2. Impact of tax credit
22.214.171.124. Renovation rate
126.96.36.199. Renovation expenditures
3.5. Sensitivity analysis
Chapter 4. Environmental degradations, fuel scarcity and women participation to labor market: Evidence from Rural India
4.3. Econometric specification
4.6.1. Forest cover in India
4.6.2. Marginal effects