{ "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": 18, "pageCount": 45, "prevPage": 17, "nextPage": 19, "slideType": "appendix_methodology", "function": "analyze_data", "density": "dense", "nDataPoints": 0, "notes": "This page appears to be a page from an academic paper or technical report, focusing on data processing and measurement error analysis.", "elementsJson": [ "paragraph", "footnote" ], "metadataConfidence": 0.9, "imagePath": null, "slideHref": "/slides/019dd923-5e88-73ef-bd59-6131f37efc2b/18", "deckHref": "/decks/019dd923-5e88-73ef-bd59-6131f37efc2b", "deckJsonHref": "/decks/019dd923-5e88-73ef-bd59-6131f37efc2b.json", "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd59-6131f37efc2b#slide-18", "components": [ { "bbox": null, "kind": "paragraph", "text": "Under the assumption that wind farm developers will do their best to minimize connection costs, we calculate the minimum distance to connect each project to the electrical grid, using a modified minimum spanning tree algorithm. Figure 4 illustrates the results for one multi-site wind power project that is spread across ten villages.", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "1cd505e5-236d-4152-8912-9fc17ffb621a", "frameworkName": null, "frameworkSlug": null }, { "bbox": null, "kind": "paragraph", "text": "This approach may underestimate connection distances, since it does not take into account topographical obstacles. The magnitude of any bias will tend to scale with the connection distance. As a result, this type of measurement error will not distort the ordering of connection distances, and therefore does not affect our empirical analysis.", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "44fee33d-0080-4763-81f9-f55c09d4462b", "frameworkName": null, "frameworkSlug": null }, { "bbox": null, "kind": "paragraph", "text": "Classical measurement error is not expected to affect our analysis. This is because it would be just as likely to lead to an erroneous BLIMP inference as an erroneous non-BLIMP inference. In our setting, a problem would arise only if our estimation method systematically underestimated the distance for CDM projects relative to non-CDM projects, or vice versa. This might occur, for example, if CDM projects are disproportionately built in places where topographical features increase the the real connection distance relative to the distance measured as the crow flies. In sensitivity analysis, we how sensitive our findings are to this possibility by deliberately inflating the distances for CDM projects relative to non-CDM projects.", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "5056fdd9-ffd1-461a-8594-99fe7246aad8", "frameworkName": null, "frameworkSlug": null }, { "bbox": null, "kind": "paragraph", "text": "greater remoteness, in and of itself, provides no particular benefit, it is easy to see that the payoff is a decreasing function of connection distance.", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "be6e3119-d678-4f24-8174-cb044878b3ff", "frameworkName": null, "frameworkSlug": null }, { "bbox": null, "kind": "paragraph", "text": "Figure 5 shows that the estimated connection distance is positively associated with project costs for the sub-sample where both are observed. The association is comparatively weak, consistent with the premise that distance to the grid affects only a small portion of overall project costs.", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "d17b5c4a-ffbf-4134-8faf-1e9f76c5f5f8", "frameworkName": null, "frameworkSlug": null }, { "bbox": null, "kind": "paragraph", "text": "We do not observe the connection distance directly and so need to estimate it. To do this, we need the geographical coordinates of the electrical substations as well as of turbines themselves. The coordinates of electrical substations were collected by Burlig et al. (2020), while the coordinates of the turbines were found using a combination of information found in the Bloomberg New Energy Finance database, the UNFCCC's CDM project database, and extensive online research, including press releases, news reports, and documents published by the Indian government. Rather than pinpointing turbines individually, we record the coordinates of the village in which the turbines are located. This was partly out of practical necessity, but also means that our findings cannot be driven by arbitrarily small differences in location. Any variation in location that we might have been able to generate at the sub-village level would likely have contained more noise than signal.", "attrs": null, "subkind": "paragraph", "toolName": null, "toolSlug": null, "confidence": null, "componentId": "d5d0b8ca-b6f1-4d7d-823f-f313a7036134", "frameworkName": null, "frameworkSlug": null }, { "bbox": null, "kind": "source-note", "text": "Appendix A provides additional details about how the raw data was processed. To avoid inflating the distances for wind farms spread across widely dispersed sites, we modified the standard algorithm. Instead of trying to connect all points in one step, we first connect the electrical substations to each other. Only then do we extend the graph to include the wind farm sites. This means that the total connection distance will include the edges that connect each cluster of turbines to its nearest substation, but avoids any edges that would be necessary to connect distant clusters. See appendix C for a more detailed description of the algorithm.", "attrs": null, "subkind": null, "toolName": null, "toolSlug": null, "confidence": null, "componentId": "6c98ec56-a786-4ffb-bc4e-1ef1bdfcf057", "frameworkName": null, "frameworkSlug": null } ], "metrics": [], "tools": [ { "name": "Analytical method", "slug": "analytical-method", "agent": null, "layer": "slide", "matchId": "d795d2f2-ae98-45bf-96ab-5293a24c1759", "evidence": "We do not observe the connection distance directly and so need to estimate it.", "confidence": 0.7 } ], "frameworks": [], "arcBeats": [], "loops": [], "imagePathAlt": null, "thumbSrc": null, "thumbSrcAlt": null, "locked": true }