Blog

Weekly Preliminary OS Climate Risk Review (4/15/21)

By April 27, 2021 No Comments

Looking Ahead While Still in Hurricane Irma’s Shadow — Despite it being two years after impact, Hurricane Irma dominated Fiscal Year (FY) 2019.

The inauspicious quote above was pulled directly from KEYS Energy Services’ (KEYS) 2019 financial report (source). Hurricane Irma had a substantial impact on the utility when it hit Key West in 2017, and the utility is still nursing its hangover following storm. KEYS does a golf clap-worthy job of disclosing its exposure to natural disasters in both its annual financial reporting and official statements. However, the utility’s effort exists in stark contrast to that of its wholesale power supplier, and this week’s non-disclosure culprit:

Florida Municipal Power Agency (All-Requirement Power Supply Project Revenue Bonds)

The Florida Municipal Power Agency (FMPA) was founded in 1978 in response to the Arab Oil Embargo that adversely impacted the U.S. energy sector in the 1970s. The energy crisis spiked the price of oil, and as a result, municipal utilities in the State of Florida formed FMPA as a mechanism for achieving economies of scale and driving down operating costs. The Agency currently works with 31 municipalities and utility boards across the State, and the 2021 FMPA bonds are secured by a pledge of revenues received via power sales contracts with 14 participating municipalities throughout Florida. Each of the participating members procure “electric power and energy, transmission, and associated services” from FMPA (POS pg. 1). In 2020, just 4 of the 14 participating munis (“Major Participants”) accounted for ~75% of aggregate ARP payments to FMPA: Kissimmee Utility Authority, City of Ocala, Beaches Energy Services of Jacksonville Beach, and KEYS Energy Services of Key West (POS pg. 4).

FMPA’s 14 ARP participants combined rank in the 99 percentile for property VaR and 97 percentile for GDP impairment nationally across the municipal utility sector by the year 2033. FMPA’s Major Participants account for ~75% of the annual revenue.

A search for the list of our favorite POS keywords and terms — hurricane, flood, climate change — leaves us empty handed. Although FMPA makes no explicit mention of the impacts of hurricane activity on its customers, events over the past 5 years tell a much different story. Hurricane Irma (2017) knocked out power to approximately 6.7 million people (65% of the population) in Florida (source), and the impact of Irma on FMPA’s power plant infrastructure and Major Participants was notable to say the least.

  • Irma “swarmed Jacksonville with the worst flooding in its 150-year history and caused an estimated $85 million worth of damage in the city,” according to reports published following the catastrophic storm (source). risQ projects that Beaches Energy Services (risQ Score 4.1; Flood risQ Score 4.9) in Jacksonville Beach could experience 39% property VaR and 67% GDP impairment as a result of a 100-year coastal flood event. Beaches Energy Services is credited for 12.30% of FMPA’s revenues in 2020.
  • Irma also forced FMPA’s 4th largest participant — KEYS Energy Services — to issue $60 million in municipal debt in order to cover costs associated with post-Irma restoration (source). Unlike FMPA, KEYS’ post-Irma official statement provides vivid detail on the threats posed by hurricane risk (source). Although FEMA provided KEYS with substantial relief funding, the event had notable impacts on the utility in subsequent years (again, from the 2019 annual report: “Irma dominated Fiscal Year (FY) 2019″). risQ projects that KEYS (risQ Score 5.0; Flood risQ Score 5.0; Hurricane risQ Score 4.9) could experience 71% property VaR and 120%* GDP impairment as a result of a 100-year coastal flood event, both ranking in the 100 percentile nationally. 
  • The St. Lucie Power Plant, primarily owned and operated by Florida Power & Light, came nail-bitingly close to shutting down its reactors in order to keep them from overheating in a power outage during Irma (source). Fourteen of FMPA’s members — including the City of Fort Pierce and Beaches Energy Services — have a power entitlement share in the St. Lucie Plant (risQ Score 4.5; Flood risQ Score 5.0). A paper authored by two University of Florida researchers in 2017 discussing the impacts of sea level rise on Florida’s energy supply noted that “decisions about what to do with the [reactor] at St. Lucie must be put at the top of our state’s to-do list” given that “decommissioning a nuclear power plant can take ten or more years” (source). FMPA participants are still heavily reliant on the power plant in St. Lucie.
  • At Irma’s peak, 53% of Kissimmee Utility Authority’s (KUA) 72,000 customers were without electricity (source). risQ projects that KUA (risQ Score 3.8; Flood risQ Score 4.3) — FMPA’s largest participant in terms of revenue — could experience 12% property VaR and 28% GDP impairment as a result of a 100-year inland flood event.

(*any value at risk percentage above 100% means that risQ projects the issuer in question could sustain cumulative losses above 100% of aggregate property or GDP value between now and the specified time horizon).

So, it kind of looks like … FMPA has substantial exposure to hurricane risk? Right? We’re bewildered but at the same time, sadly, unsurprised that the Agency fails to mention (even out of courtesy) its historical and future exposure to hurricane activity. And, unfortunately for FMPA, the risk of future hurricanes is only intensifying as climate change alters the variables that drive these storms.

The following table provides a view of the changes in hurricane probabilities within the 4 counties encompassing the Major Participants — KEYS Energy Services (Monroe County), City of Ocala (Marion County), Kissimmee Utility Authority (Osceola County), and Beaches Energy Services (Duval County). Let’s drill down into some climate-conditioned views of KEYS’ hurricane risk over the coming decade. Using climate scenario RCP 4.5, the annual probability of a Cat 1 or greater hurricane hitting Monroe County is 11.3% in 2021 and 13.4% in 2033, an increase of 19%. The increase in annual likelihood becomes starker when we look at the RCP 8.5 scenario, in which emissions continue ‘business-as-usual’. In this scenario, Monroe’s annual probability of a Cat 5 hurricane is 0.77% in 2021, and by 2033 it increases to 2%, a 160% increase in probability.

We can also extract some interesting macro-observations from these projections of increasing hurricane activity. While the exact increases in annual probability vary by county, we see that by 2033 (under the conservative RCP 4.5 scenario), the annual likelihood of a) Cat 2 or greater hurricane increases by ~20-25%, b) Cat 3 or greater increases by approximately 32%, and c) Cat 5 by approximately 75%. 

By aggregating these hurricane probabilities over the 13 year term of this series of FMPA bonds dated out to 2033, we can achieve a fully fleshed out view of the Agency’s hurricane exposure.

The above table shows the likelihood of a hurricane at least meeting each category threshold by 2033. We see that in Monroe County, the probability of a Cat 1 or greater storm by 2033 is 82%, and the probability of a Cat 5 storm is 11% by 2033, even under the conservative RCP 4.5 scenario assumption. risQ is projecting that if a Cat 5 hurricane hit KEYS today, the ensuing storm surge and extreme wind would result in 77% and 19% of property value losses, respectively.

KEYS 2033 outlook appears, well, grim. And therefore, so does FMPA’s prospect of consistently and sustainably generating revenue from its risQiest Major Participant. Although the Agency is geographically diversified throughout the State of Florida, there appear to be kernels of underlying hurricane and flood risk throughout the basket of FMPA participants that will increasingly disrupt power generation and transmission, and economic activity. We’ve become accustomed to finding boilerplate climate risk disclosures in the official statements of Florida municipalities (“The State is naturally susceptible to the effects of extreme weather and natural disasters…”), but in our eyes, FMPA has achieved a new low. Cybersecurity and COVID-19 each got a plug in the POS, but not Climate Change? We find that personally offensive, and so should the investor eventually holding the FMPA bag.

Honorable Mentions

Tangipahoa Parish School District, LA (risQ Score 3.1; Flood risQ Score 3.8)

A quick paragraph on page 15 of the OS succinctly summarizes that there is climate risk and flood risk in particular. Within that, there is acknowledgement of climate change assuming scientific projections “are correct”, that the school district’s insurance may not be sufficient, and that the population may not choose to live there anymore in the aftermath of events. 

There isn’t any discussion of efforts to mitigate risk in the document, which implies for all the potential risk that’s minimally recognized, there isn’t anything to point to into terms of risk mitigation, for those that want some “G” in their ESG. We can’t find evidence elsewhere either. While the parish is part of the NFIP Community Rating Survey, it’s only a 9 on that 1-10 scale where 1 is best. A 9 is maybe marginally better than filling in your name at the top of an exam and worse – sometimes significantly – than the surrounding high risk parishes in the state. As a result, only 5% discounts on flood insurance are available to residents, a problem when less than 20% of the flood risk to residential property in the parish is currently insured.

Clearly there is enough to be concerned about though from past events. Parish residents have lodged NFIP claims/capita/year of over $30,000, in the worst 3% of counties nationally. Buy-outs of serially flooding properties have occurred, but only at the 85th percentile level nationally, so the parish again lags in taking action versus the damage that has been inflicted. Beyond all this, Zillow data shows that growth of residential property’s market value of property in the parish is in the bottom 8% of counties nationally, but the assessed property values have far outstripped that growth. How much questionably ad valorem tax burden, insurance cost burden and/or flood repair cost burden can the population take? Keep in mind that 22% of the population is below the poverty line  –  that’s in the worst 12% nationally – and 43% of renters are categorized as “stressed”, in the worst 8% nationally. Something’s gotta give, and climate risk appears to be a force that the parish isn’t doing a whole bunch to stop.

The School Board of Orange County, FL (risQ Score 3.1; Flood risQ Score 2.4; Hurricane risQ Score 2.4)

A single 7 ½ line paragraph on “Climate Change and Natural Disasters” in this one down on page 79, with the usual odor of copy/paste from some other document without any actual thought as to the actual risks and what to do about them. Last time we checked, Orange County is inland so mentioning sea level rise even in the minimal language that is there is nonsensical. They could have actually done themselves some favors too. After all, both Orange County and the City of Orlando are in the NFIP Community Rating Survey, and with respectable scores of 5 and 6, respectively. A rating of 5 means that there has likely been at least some transition from talk to action. That said, most of the residential flood risk in the district remains uninsured.

Another maybe more noteworthy section is the “Property Insurance” language that comes immediately after the aforementioned faux climate change disclosure. Notably, they acknowledge that hurricane and flood insurance premiums have gone up “dramatically” in the last few years and they are buying lower levels of insurance as a result. This leaves the issuer much more exposed to climate risks than ever before, at a time when climate risks are only going to increase. This is only the beginning of the insurance cost escalation you should expect in the coming years so the exposed risk to the issuer, and therefore the holders of these bonds will only climb with it. This $108 million in total par value is backloaded to 2032, the final year, so you will have 11 years of insurance cost increases, uninsured exposure increases and overarching increase in material risk to navigate. Buyer beware.

City and County of San Francisco, CA (risQ Score 0.3; Flood risQ Score 1.2)

Why highlight an issuer with such relatively low risk? Check out the extent of climate risk discussion in this OS versus what you get from issuers (including the likes of the above) who have much more problematic climate risks.

There are actually two parts to this story. There is a full two pages (p25-27) of discussion of “Climate Change, Risk of Sea Levels Rise and Flooding Damage providing a full account of the risks, the strategy and the actions for mitigating those risks. Even the opening paragraph sets the appropriate tone stating that climate change effects are that “sea levels will rise, extreme temperatures will become more common, and extreme weather events will become more frequent” (emphasis is ours). Just a touch different than the nebulous and subtly dismissive language we see from other issuers who even take the time to mention the climate risk and climate change are even a thing. The remainder of the two pages then lays out the specific risks and projects being pursued to mitigate those risks. In addition, the city’s lawsuit against oil companies is referenced, which most assuredly covers the ESG spectrum in the broadest context possible. This is all gold standard and, again, from an issuer that has far less reason to worry than many for quite some time.

Later, the “Other Events” section from page 28-29 spends the vast majority of the time discussing wildfire risk, and especially to San Francisco’s water supply and the Hetch Hetchy project that currently provides 85% of the city’s drinking water. This is for good reason too given the history that is discussed in the OS and that our own analysis of that area – captured as the Twain Harte – Tuolumne City CCD – shows a risQ Score 3.4 and, more specifically a Wildfire risQ Score of 4.2. This problem is not going away and is more of a threat in the short and medium term to the city than anything discussed earlier. The risk has clearly been disclosed and that’s better than most, again. However, this is also where the OS actually falls short in two ways. Firstly, it has decoupled climate change from wildfire risk which we know is wrong and, secondly, the OS has no discussion of efforts to mitigate the risk. Instead, the last paragraph of this section essentially states that they’ll deal with the financial exposure and let the cards fall where they may.

Again, this is about as good as climate risk disclosure gets from an issuer, but even the City and County of San Francisco has climate disclosure development needs to address.

Leave a Reply