{ "docId": "019dd923-5e88-73ef-bd59-6131f37efc2b", "docSlug": "71b5b657f034a972", "documentTitle": "ey net zero centre carbon offset publication 20220530", "authorId": "MorganStanley", "authorName": "Grantham Research Institute", "documentKindSlug": "consulting-deck", "documentKindLabel": "Consulting deck", "sourceTypeSlug": "equity_research", "sourceTypeLabel": "Equity research", "presentationDate": null, "orientation": "portrait", "aspectRatio": 0.707, "pageNumber": 27, "pageCount": 45, "prevPage": 26, "nextPage": 28, "slideType": "appendix_methodology", "function": "diagnose", "density": "dense", "nDataPoints": 0, "notes": "The text refers to 'Table 1' and specific rows (8-13) which are not visible on this page. The content is academic/technical analysis of methodology.", "elementsJson": [ "paragraph" ], "metadataConfidence": 0.9, "imagePath": null, "slideHref": "/slides/019dd923-5e88-73ef-bd59-6131f37efc2b/27", "deckHref": "/decks/019dd923-5e88-73ef-bd59-6131f37efc2b", "deckJsonHref": "/decks/019dd923-5e88-73ef-bd59-6131f37efc2b.json", "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd59-6131f37efc2b#slide-27", "components": [ { "bbox": { "h": 0.085, "w": 0.766, "x": 0.117, "y": 0.285 }, "kind": "paragraph", "text": "Another potential omitted variable is the number of sites a project is spread across. The number of sites could affect the cost to build and to connect a wind farm. Row (9) matches projects on state, vintage, and number of sites. As before, this makes it harder to identify BLIMPs, but in relative terms, the CDM still performs poorly when compared to a lottery.", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "1ea5583a-5aee-4a15-9448-71e75a2d1e8f", "frameworkName": null, "frameworkSlug": null }, { "bbox": { "h": 0.185, "w": 0.766, "x": 0.117, "y": 0.519 }, "kind": "paragraph", "text": "Arguments about omitted policy variables can be extended down to the district-level, and in the limit, down to the village-level. Perhaps there is systematic variation in policies (or factor prices) between districts or villages within the same state, which affect costs. In rows (11) and (12) we match wind farms in the same districts and the same villages. It now becomes much more difficult to find comparable wind farms and to spot BLIMPs. Matching at the village-year level has the side-effect of also eliminating variation in connection distances and capacity factors, so BLIMPs are only identified based on size differences. Even so, measured either in BLIMP capacity or offsets, the CDM still performs poorly relative to a lottery.", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "46741036-66a3-4a5e-8ac7-719454377b31", "frameworkName": null, "frameworkSlug": null }, { "bbox": { "h": 0.185, "w": 0.766, "x": 0.117, "y": 0.093 }, "kind": "paragraph", "text": "turers differ in their pricing, quality, technical specifications, ability to supply large developments, maintenance costs, etc., we could imagine that the choice of manufacturer affects both the profitability of a wind farm and features like its capacity or capacity factor. In row (8) of Table 1, we therefore require projects to use turbines from the same manufacturer in order to allow one of them to be classified as a BLIMP. Unsurprisingly, BLIMPs are much rarer occurrences if we require that wind farms are built in the same year, in the same state, and use turbines from the same manufacturer. However, taking into account that the same difficulty extends to counterfactual CDM assignments, we see that the CDM still performs quite poorly compared to a lottery (except for the number of BLIMPs).", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "4f4d6e8c-3b69-43d6-8e87-032206d5a5f5", "frameworkName": null, "frameworkSlug": null }, { "bbox": { "h": 0.02, "w": 0.02, "x": 0.485, "y": 0.96 }, "kind": "paragraph", "text": "25", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "805c066e-e4f9-4e07-b4e8-aa5cd6242b55", "frameworkName": null, "frameworkSlug": null }, { "bbox": { "h": 0.135, "w": 0.766, "x": 0.117, "y": 0.377 }, "kind": "paragraph", "text": "Another possibility is that projects above 5MW are somehow different than projects below 5MW. As discussed, the Generation-Based Incentive specifically supports wind farm developments that exceed 5MW in capacity. If the policy is in place to compensate for some unobserved cost of scale, such as the loss of support from local governments then our results may be biased. In row (10) we match wind farms on the year of construction, the state, and on whether it exceeded the 5MW capacity threshold. This improves outcomes for the CDM, but a lottery would still outperform the CDM most of the time.", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "bce4825f-3a7d-486a-87c2-7b3a986217e3", "frameworkName": null, "frameworkSlug": null }, { "bbox": { "h": 0.215, "w": 0.766, "x": 0.117, "y": 0.711 }, "kind": "paragraph", "text": "It is also possible that differences in project developer are more important than a feature of the wind farm, or a function of its location. Project developers may differ systematically in their engagement with local stakeholders, in their adherence to local regulations, and in their ability to develop wind projects more broadly. The CDM may quite reasonably support only projects put forward by reputable developers. If so, our results may be driven by the fact reputation is only observed by the CDM. Row (13) shows that the CDM's performance relative to a lottery improves when we limit our sample to projects built by developers with at least one CDM-registered wind farm. However, even within this restricted sample, the lottery outperforms the CDM ≈ 700 times out of 1,000. While there might be some merit in the \"developer quality\"-hypothesis, it explains at best a small part of the CDM's apparently poor performance.", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "d9c1353e-2874-4f9d-8246-e2a0f8e66c18", "frameworkName": null, "frameworkSlug": null } ], "metrics": [], "tools": [ { "name": "Logical chain", "slug": "logical-chain", "agent": null, "layer": "slide", "matchId": "53f4aafd-e280-450d-a2e0-c881d7db49aa", "evidence": "Arguments about omitted policy variables can be extended down to the district-level, and in the limit, down to the village-level.", "confidence": 0.6 } ], "frameworks": [], "arcBeats": [], "loops": [], "imagePathAlt": null, "thumbSrc": null, "thumbSrcAlt": null, "locked": true }