Carbon offsets = Real reductions?

There is increasing use of opt-in regulation to encourage environmental protection at minimal cost to economic growth, particularly in developing countries. But, how do we determine which programs are effective? Are carbon credits more than hot air?

A growing number of environmental protection policies focus on rewarding good behaviour rather than punishing bad behaviour. One important example is carbon offsets, whereby one party (say, a manufacturer in India) voluntarily agrees to reduce emissions for payment from another party (say, a frequent flier who buys a carbon offset from an airline). It’s not always obvious, however, to determine how much to pay those who reduce emissions. We want to reward them relative to what their level of emissions would have been but for the intervention.

To allocate the right number of credits, regulators need to estimate what economists call a “counter-factual” emissions trajectory. Specifically, they need to determine that a firm would not have done the project anyway, and, if so, decide exactly how many credits to award. Offset projects typically improve environmental outcomes by displacing dirtier processes, so the number of credits awarded depends critically on what is displaced (coal, natural gas, or other renewables). And total emissions depend on expected growth, which can be difficult to project.

The difficulty of this task is compounded by the fact that firms have incentive to overstate how dirty they would have been. There is also concern that participating firms are not necessarily those with the lowest-cost projects but those who expect to benefit from over-generous baselines.

We looked to the largest global carbon offset program, the Clean Development Mechanism (CDM), for insights. One of three flexibility mechanisms in the Kyoto Protocol, the CDM provides incentives for greenhouse gas emission reductions in industrializing countries like China, India, Brazil, and Mexico. Other countries and firms in those countries can meet part of their Kyoto commitments using CDM Certified Emission Reductions (CERs) in the EU Emissions Trading Scheme. Between 2006 and 2012, the CDM credited 1.5 billion tons of carbon reductions.

Unlike previous studies, we don’t just look at participant behaviour. We match projects from the database of all projects submitted into a large accounting database of manufacturing firms. We can, thereby, observe firms that apply, those who don’t, and those rejected, both over the course of the program, as well as before the program was announced. We match all applicants to similar firms that did not apply to evaluate both who offers to reduce emissions and by how much relative to their peers.

India, the second-largest source of CERs and host of over 1400 projects, provides a rich foundation for our investigation. We focus on the manufacturing sector because it is an essential and emissions-intensive component of India's economy. And unlike the case with most renewable generation projects, manufacturing firms can be observed before they decide to apply to the program.

Who supplies emissions reductions to the CDM?

Applicants are on average larger and, therefore, better able to afford the high fixed costs associated with participation.

Firms with low cost of reducing emissions are ideal applicants.

CDM applicants and participants are more likely to be expanding capacity, consistent with the view that abatement is cheaper at the investment stage than as a retrofit.

But we do not see that applicants otherwise come disproportionately from the pool of firms that would receive the highest credit for displaced emissions, due say, to having old technology, or being in a region with high grid emissions. Firms with a low cost of capital are not particularly likely to apply. Human capital intangibles, like managerial expertise, which have been shown to play a strong role in the decision to export, do not appear to directly drive participation above and beyond a correlation with firm size.

We assign fuel-based emissions using firm-level input data on physical quantities and expenditures disaggregated by fuel type, and attribute additional process-based emissions for cement firms based on limestone inputs. Encouragingly, we find no evidence that firms on trends of decreasing emissions are more likely to apply to supply carbon offsets, pointing to a lack of adverse selection by firms hopeful of being overly compensated.

Do participants really reduce emissions?

We then examine the impact of CDM participation on fuel-related carbon emissions during the program, relative to pre-CDM emissions. Controlling directly for pre-treatment trends and characteristics at the firm level, we find that, in the aggregate, participants reduce their emissions intensity of output faster than non-participants and, to a lesser extent, also have emissions profiles that grow at a slower rate. The effects are larger in the later years.

There is strong heterogeneity across project types. Relative emissions drop thanks to projects that fund co-generation and exporting excess energy to the grid and, to a lesser extent, energy efficiency. Firms with projects that reduce emissions per unit output by blending additives into the final product, and firms that get paid to switch fuels, in the aggregate, experience either increases in emissions or no net benefit. It appears that many firms, even those who are not receiving CDM credits, switch from coal to natural gas.

We then compare the emissions reductions observed to the number of credits received. Although CDM participants are reducing emissions relative to other firms, they are doing so much less than the volume of credits received.

So, what should we do?

What is most interesting about our analysis is that there is no reason why it has to be done after the fact. Regulators should seriously consider reforming the CDM to credit firms based on their performance relative to similar non-participants. Applicants can be matched to non-participants anonymously ahead of time, and emissions credits allocated based on actual performance relative to this counterfactual group. Firms that want to be guaranteed a certain number of credits up-front can contract with intermediaries who take on the risk. The transaction costs in such an offsets system would be significantly lower, allowing more firms to apply. And as long as participants don’t affect non-participant firms through general equilibrium factors such as fuel prices or diverted access to finance, the environmental integrity of the system would be ensured.