Tesla Issues Statement On Paint Shop Problems

Tesla's statement boils down to two fundamental claims: that this entire situation is the result of its proactive transparency and engagement with BAAQMD, and that none of the deviations from permitted conditions resulted in emissions exceeding the factory's allowable levels or "adverse community or environmental impacts." Though InsideEVs is apparently willing to take these claims at face value, we can think of several reasons why Tesla didn't provide this statement in response to our original reporting and having spent hours poring over the relevant documents and spoken with sources familiar with the situation we aren't convinced that it is the last word on the matter.

To begin with, it's important to acknowledge that Tesla doesn't refute any of the specific factual claims in The Drive's reporting. Though their statement attempts to recast the past 18 months of elevated deviations and violations as having no environmental or community impact, the unique nature of these events and both BAAQMD and Tesla's response to them proves that the situation is indeed serious. The uniqueness of this situation is in fact unwittingly proven by InsideEVs' comparison of it to a 2003 violation by New United Motor Manufacturing Inc (NUMMI) at the same facility.

That single 2003 incident reflects a far more typical air quality violation at an automotive paint shop: a routine test of a single VOC emission abatement device in October 2003 suddenly showed that it was operating at 46% efficiency instead of the required 90% efficiency. NUMMI received those test results in November, ordered another test by another firm in December and when that test confirmed the results NUMMI replaced the faulty parts by the end of January. NUMMI paid fines for the excess emissions between the discovery of the unexpected problem and its resolution, but crucially it also went before the District's Hearing Board to request a variance for its period of noncompliance because simply shutting down production would not have put it into compliance because the regulation it was violating was based on abatement efficiency and not simply the level of emissions (which did not exceed the limit for the bumper line, even without proper abatement).

This event was a typical violation in that it was an isolated incident caused by the unexpected failure of a single piece of equipment that was repaired promptly. More importantly, because NUMMI requested a variance requiring a hearing before BAAQMD's Hearing Board the public has access to a comprehensive report on the situation [PDF] that shows the incident was unexpected, isolated and handled promptly, completely and with true public transparency. All of these characteristics fundamentally differentiate this incident from the current situation at the Tesla factory over the last 18 months, making InsideEV's comparison between the two situations more reason for concern than reason to dismiss the matter.

For the 19 violations at Tesla's factory to have been comparable to the NUMMI violation, there would have to have been a statistically improbable number of random and unexpected equipment failures as well as evidence that Tesla promptly returned them to compliance, provided a fully transparent accounting of the situation and sought variances before the Hearing Board for the period of noncompliance. This is clearly not what happened, as is obvious to anyone who reads the publicly-available documentation of the situation, or The Drive's narrative account of the problems. But because this is a complex situation and because neither Tesla nor BAAQMD is being as transparent about the situation as the public deserves, not to mention Tesla and InsideEV's efforts to muddy the waters, it falls on us to present this as simply as possible.

Were Tesla's violations all isolated instances of unexpected equipment failure? 

No, the deviations and violations cover a range of equipment and emissions sources across the factory and in many cases they involve the failure to obtain proper permits before installing, modifying and/or operating equipment. This pattern of not waiting for regulator approval before taking actions that alter conditions from those laid out in the engineering analysis for Tesla's current permit is fundamentally different from discovering an isolated problem and fixing it as quickly as possible, and it reflects a more fundamental disregard for air quality regulation.

Though there are reported deviations that do seem to involve the sudden failure of abatement equipment, there are others where Tesla simply deactivated E-Scrub abatement equipment due to fires that were at least partially caused by poor maintenance practices. The shutdown of E-Scrub devices in turn led to a buildup of primer and clearcoat in ductwork and sprinklers, damage to regenerative thermal oxidizers (RTOs, an abatement device that burns off 90% of VOCs when properly functioning), source tests showing POC emissions beyond permitted levels from one of those RTOs.

Did Tesla appropriately respond to noncompliance? 

Based on currently-available information, the answer appears to be no. In contrast to the NUMMI violation cited by InsideEVs, Tesla responded to many of these noncompliances by installing unpermitted equipment and making other unpermitted modifications rather than simply fixing or replacing broken equipment. More importantly, before April of this year Tesla did not go before BAAQMD's Hearing Board to formally request variances for these problems as NUMMI did. These formal requests for variance represent the appropriate way for a company to "make a calculated decision" (to use InsideEV's characterization of the situation) to not comply with existing permits, yet Tesla did not pursue them until after BAAQMD stepped up its enforcement efforts.

A review of Hearing Board records shows that Tesla has only requested a single variance since the start of 2018. On April 17, 2019 Tesla sought and obtained an interim variance for RTO A-1002 [PDF} for the period between March 28, 2019 and June 24 2019 which allowed it to operate at 1275 degrees Fahrenheit rather than the usual temperature of 1400 degrees F. This variance was granted over BAAQMD's objections because Tesla claimed that A-1002 was damaged and that operating it at the normal temperature could cause it to shut down, creating more emissions than it would allow at the lower level. 

The fact that Tesla has now requested a variance for A-1002's final few weeks of operation before it is replaced by an entirely new RTO (A-1007) suggests that its behavior is only changing as a result of BAAQMD's higher level of scrutiny and tougher enforcement actions in the last year. A-1002 has been a problem since its NOx levels were found to be out of compliance in 2014, and after that problem was supposed to be fixed per the schedule of compliance [PDF] Tesla submitted two separate applications to replace its burner in 2016 (applications #26899 and #27587) and again in January 2019 (#29741) suggesting problems continued beyond the 4.5 days when it broke down completely in October 2018, allowing unabated emissions from an e-coat oven. Critically, Tesla failed to maintain records for A-1002 from 2016 until October 2018, and its annual source test in December 2018 failed to capture all of A-1002's emissions parameters due to "a miscommunication between Tesla and its source test contractor." 

Had Tesla taken the action that NUMMI did (and which they have subsequently done for A-1002) and sought variances for each of their instances of noncompliance, they could have thoroughly explained the nature of its problems, made the case for its subsequent response, received the District's blessings for its actions and provided a clear public record explaining the entire situation. The fact that this did not take place prior to BAAQMD's increased engagement and enforcement not only shows that Tesla hasn't responded to this situation in a formal and above-board manner, but it also raises questions about the extent to which its deviations even disclose the full extent of its problems. 

Was Tesla transparent about its problems and noncompliance?

Tesla's statement argues that all permit deviations were self-reported, which in InsideEVs' estimation reflects the company's good-faith efforts in this matter. It is true that the semi-annual report of deviations that The Drive has relied on for its reporting is generated by Tesla, but those semi-annual disclosures are required by law and also far from the extent of Tesla's reporting requirements. That an ostensibly environmentalist outlet like InsideEVs believes that it would be sufficient for Tesla to report deviations from permit conditions as much as six months after they have taken place beggars belief.

Indeed, a 2004 bulletin [PDF] by BAAQMD's compliance and enforcement division reminds Title V permit facility operators like Tesla  that "all instances of noncompliance with the permit shall be reported in writing to the District’s Compliance and Enforcement Division within 10 calendar days of the discovery of the incident. Within 30 calendar days of the discovery of any incident of non-compliance, the facility shall submit a written report including the probable cause of non-compliance and any corrective or preventative actions." 

The fact is, we still don't know precisely how and when BAAQMD learned of each of the deviations Tesla has reported in its semi-annual reports and The Drive is currently requesting the necessary documents to clarify this question. It is clear from the semi-annual reports alone, however, that at least some of the deviations were discovered by inspectors and source tests suggesting they were not all promptly and proactively reported by Tesla. And because many of the deviations are the result of Tesla simply ignoring the permitting process in order to make unapproved modifications, the company clearly hasn't been "working closely with the agency." Moreover, the lack of proper record-keeping and reporting at some of the sources and abatement devices at the center of some of the most troubling deviations show that basic transparency measures weren't followed.

Had Tesla sought and obtained variances from the District's Hearing Board it would have provided a complete and publicly-available accounting of the entire situation as well as proving that it was making a sustained good-faith effort at compliance. As District Regulation 1-1-402 [PDF] makes clear, had Tesla applied for a variance the District would not have been able to issue notices of violation between the date of application and the date of the Board's hearing on the matter. So one of the reasons Tesla has accumulated 19 notices of violation over the last 18 months is the fact that it was neither fully transparent nor making the full good-faith effort required to remain in compliance with regulations.

Did these failures result in an environmental or community impact?

It's quite surprising that Tesla would unequivocally state that there has been no environmental or community impact resulting from its actions over the past 18 months, given that only a regulator is in the position to make that determination. In its most recent statement to The Drive, BAAQMD has made it clear that it is continuing to investigate a number of the deviations Tesla has reported. Without knowing the details of the 19 violations that have already resulted or how many could still result from the deviations still under investigation, let alone the results of source and area testing/monitoring, the community impact is still unclear and The Drive has not alleged otherwise.

Tesla's claim that it "we have not exceeded the Fremont factory's total allowable emissions, and have in fact been well below allowable levels" isn't completely surprising given that its current permit was calculated based on the projection that it would build 520,000 units per year in 2018. Given that Tesla built 77,100 vehicles at Fremont in the first quarter, for an annualized production rate of just 308,400 vehicles, one certainly hopes that they have not exceeded emissions permitted for 520,000 vehicles per year. Had they done so, they would be emitting nearly twice the level of pollutants per car produced envisioned in their permit, implying an almost unimaginable failure of either permitting or abatement.

However, as the NUMMI incident highlighted by InsideEVs shows, a facility can remain under its permitted emissions level and still be fined for being out of compliance with its permit. More importantly, because Tesla has clearly not been working with the District until very recently, requesting variances, or keeping proper records at sources and abatement devices that have been shut off or failed, its unsupported assertion that there has been no environmental or community impact is not particularly credible. 

The reason that the Authorities To Construct and Permits To Operate that Tesla has simply been ignoring are so important is that an air quality permit is like a complex mechanical machine, and when one part gets changed the rest of the parts no longer interact the same way or produce the same outcomes. This is why a single modification, like de-activating E-Scrub machines produce results like damage to the RTO, which in turn affects destructability of POCs from other sources. The reason BAAQMD takes so long to issue new permits is that it needs to understand all the possible upstream and downstream results of that single change, and with as many modifications as Tesla has reported making at its factory there's no way it (or possibly even BAAQMD) can say with total confidence what all the collective impacts add up to.

So, what now?​

If you're still reading this, you should have a sense at how complex this situation is but also of how clear it is that Tesla routinely failed to operate its factory in compliance with clear air rules. It is heartening to see that BAAQMD is taking a more active role in the situation, and that Tesla has finally started to formally request variances for faulty and damaged equipment, but that is not a reason to simply pretend as if the extraordinary events of the past 18 months never happened.

We are continuing to request the records required to more fully understand what has happened, and what both Tesla's and BAAQMD's responsibility for this situation. We have discovered irregularities in the permitting record that we continue to investigate, and we are also looking into what connection (if any) these events have to the recent surge in reports of weaker-than-usual paint on delivered Teslas. We are also examining what (if any) role air quality permitting issues have played in Tesla's inability to achieve its promised 500,000 units/year production level at Fremont and what limitations the paint shop specifically places on the facility's production rate.

These are complex and challenging questions, and we invite other media outlets to join our effort. We offer any assistance and collaboration we can provide, and will continue to post source documents whenever possible to aid in other investigations. We also call on BAAQMD to proactively disclose as much information about this situation as their policies allow.