jfan
Member-
Posts
430 -
Joined
-
Last visited
-
Days Won
2
jfan last won the day on March 29 2023
jfan had the most liked content!
About jfan
- Currently Viewing Forum: General Discussion
Recent Profile Visitors
2,294 profile views
jfan's Achievements
-
Thanks @Cigarbutt. Your post was very helpful. I came across a blog (only glanced at it) discussing converting calendar year to accident year which I've linked below: Calculating an insurer’s accident year and calendar year reserves | kitchensinkinvestor (wordpress.com) At my stage (can only trust but not yet verify), I can only take Prem's words that they are underwriting conservative and consistently with the goal to survive. From a top down view, the following "soft" items suggest this is taking place: 1) Not letting go of employees when there is a soft market 2) Not acquiring turnaround insurance operations 3) allowing insurance subs to operate in a decentralized manner to flatten barriers to information flow 4) facilitating fun competitions to innovate between organizations, to share best practices 5) tenure of insurance presidents in the organization 6) cautiousness in entering new territories ie cybersecurity (small exposure, recognizing Cat-like behaviours) 7) the obvious data trends for combined ratio/loss ratio, net premium written:surplus ratios, reserve triangle patterns Last but not least, the kindness of the CoBF board to scrutinize Fairfax
-
https://www.cbc.ca/news/politics/federal-budget-2024-main-1.7175052 This might have an impact on Fairfax....
-
Thanks @Cigarbutt and @Tommm50 for your replies and insight. The idea of writing policies in a soft market with just adequate pricing (with the expectation of some adverse loss developments) when there are good opportunities to use those premiums to invest in their portfolio for more attractive returns was helpful. Spending the morning watching videos on insurance triangles, paid/case reserves/incurred but no reported, basic methodologies to estimate of IBNR to derive loss reserves (case reserves + IBNR), and how to interpret loss development trends was particularly stimulating. Looking back at the 2 tables that I posted. A few amateur observations were made: 1) looking down the columns for each calendar year: a) It appears that ~ 70% of cumulative paid losses occur within 6 years (6-year cumulative payment divided by 6-year re-estimated reserves) b) Unfortunately, they don't report the entire table in 2023 AR, in the 2022 AR, after 8-9 years, the proportion of cumulative payments is mid-to-high 70% of re-estimated reserves). This suggests to me a fairly long-lag of actual payout to claimants. I'd assume this is because of the longer-tail nature of their specialty insurance focus. c) Eyeballing the re-estimated reserves down each calendar year, there is little variability year-to-year as time goes by, within ~ $500 M at most up or down (in the range of $100 - 300 M most the time). Coupling what was said earlier about the soft market, it seems that there is not much pricing power, or too much social/judicial/cost inflation to be able to adjust reserves in a favorable manner. 2) Looking at the diagonals for cumulative payments (gross table, not net) a) The sum of diagonal (latest) minus the sum of the diagonal (previous year) = the payment made for losses in the most current calendar year. Using the 2022 AR, they paid out $25.5 B and had a final year end reserve of $38.3 B (ie ~ 66% or for every $1 of reserves set aside, they paid ~ $0.66 that year). Using the 2023 AR (which of there are only 6 years), they paid out $28.1 B and had a final year end reserve of $41.2 B (ie ~ 68%). This seems to me that their baseline underwriting assumptions are that 2/3 of their reserves will be paid out ~ 6 - 9 years. I have not compared to their prior years nor other insurance companies, but is this in the opinion of the experts here, conservative? I would imagine that if their underwriting deteriorated and they were not re-estimating their reserves properly, then these ratios should deviate from their baseline over time. I think this might be more useful than just looking at the reported magnitude of adverse/favorable loss reserve developments which is more retrospective than prospective.
-
I was just looking at their reserve triangles and it seems they are reporting a number of years of adverse developments (gross undiscounted calendar numbers). I've attached 2022's and 2023's triangle. Just wondering if someone can provide me so insight and whether this is something to watch more carefully (percentage-wise it doesn't seem to be very large relative to their policy liabilities)
-
Found a couple of interesting resources looking at how the value ($) of stocks traded as a % of a countries' GDP. World Development Indicators | DataBank (worldbank.org) Animation: Stock Market vs. GDP Share, by Country (1900-2022) (visualcapitalist.com) - This one had an interesting visual over >120 years of data. The US stock market has roughly had a stock market share 2x their GDP share since 1930s. It fluctuates between 2-3x then compresses. Japan at one point had a stock market share 4x their GDP before collapsing. I know this is a very top down view, but seems like US dominance of the stock market has been quite persistent and China is still quite nascent in its development and is only 0.25x their GDP (if you believe their GDP numbers?).
-
It seems that the HK exchange and clearing Ltd is a holding company that owns the exchange and a clearinghouse as subsidiaries. It's unclear if they use their own clearinghouse for their own corporate funds or someone else's. They have a significant amount of corporate cash that they use to fully manage internally. In 2016-2017, they started allocating to external fund managers. They call this their "external portfolio" or "collective investment scheme". They started with 15 funds but at one point had ~ 32 managers. They are at ~30 right now. They also have clearinghouse funds that they manage using liquid, short-term investment grade instruments. This was found in their annual report. I've included their most recent YE income statement, and their investment revenue is not an insignificant portion of their overall revenues. Hence, the question of conflict-of-interest and "insider"-like information regarding fund flows being an exchange and clearinghouse itself.
-
I was just looking at the Hong Kong stock exchange and came across something odd. They have a significant amount of cash and decided instead of providing shareholders with special dividends, they opted to invest this cash in an investment portfolio (70% bonds, 30% equities that are short-dated and liquid). This seems like a conflict of interest, ie visibility in the data coupled with investing their own cash coffers. Is there something else I'm missing here? Should the regulators have stopped this? Is this common practice among exchanges?
-
The Silk Road admin sentenced to 2 life sentences. The Alpha-bay admin committed suicide in a Thai detention center. The Welcome to Video admin got 18 months in a South Korean jail. The BTC-e admin sentenced to 25 years. Don't do illicit things on the blockchain.
-
Some interesting tidbits from "Tracers in the Dark" 1) 2021, as per Chainalysis - $14 billion of criminal transactions - 0.15% of crypto transactions are illicit - 400 employees - 600 customers (law enforcement, tax agencies, financial institutions, exchange companies) - Chainalysis' valuation $8.6 billion - Their competitor - Elliptic 2) During the takedown of Welcome to Video, a South Korean run child abuse darkweb site, it was discovered that a US homeland security border agent was an active participant on this site. He was abusing his girlfriend's daughter and posing as a moderator for the site to steal login/passwords of other users to obtain videos. He plead not guilty because he argued that his BTC transactions were surveilled without his consent and violated his privacy. Courts turned down that argument because all transactions on the blockchain are public and therefore not subject to privacy laws. 3) Ransomware gangs are trending to the use of Monero and Z-cash where "privacy" is much harder to break using more traditional blockchain cluster analysis techniques. This got me thinking about Jason Lowery's Softwar book and how BTC given its transparency and proof-of-work protocol, could help impose a real-world cost to ransomware gangs to access centralized database system.
-
Chainalysis has a plethora of information on its website. This is their latest report. I like this little snippet, unfortunately, I couldn't find a publicly traded company in this industry. The 2024 Crypto Crime Report.pdf
-
Couple of old academic articles 1) On the traceability of BTC due to its open ledger system (A Fistful of Bitcoin) - misconception that large illicit activities are best done with BTC 2) On the fallacy of BTC immutability and a different form of governance/risk of 51% attack - BTC was smaller at the time of this article, but highlights that governance of BTC just in a different form Recommend reading Tracers in the Dark on the history of Silk Road, Mt Gox, Alpha-Bay, BTCe, Chainlysis and FBI investigations/corruption. A Fistful of Bitcoin.pdf The_Economics_of_Bitcoin_Mining,_or_Bitcoin_in_the_Presence_of_Adversaries.pdf
-
4.4% at cost 8.1% at mark-to-market 13.7% cash
-
lol...already running one and a nerdminer. It would interesting to see if people start contributing to the hash rate in a small scale non-economic manner. Better than spending money buying lottery tickets weekly.
-
The blockchain war book is particularly interesting and ties in many of the debates we have here: 1) payment system vs non-sovereign money 2) A centralized system of power vs a decentralized system of governance Book describes 2 camps within the bitcoin ecosystem, big blockers and small blockers. This debate arose because of the amount of data that each block can contain is 1 Mb which causes a limitation in scaling the BTC network as a payment system. Interestingly, the big blockers comprised mostly of the China miners/mining equipment developers/big mining pools/many American-based crypto exchanges (Coinbase), and those representing big business focused on a non-Visa payment system. The small blockers comprised primarily of developers and those that wished to create a new decentralized money (their ethos was open-source, network community consensus on the underlying rules). The author describes the big blockers having a shorter time frame, believed that amount of hashing power represented the degree of influence they should have, and their understanding of the computer science was less skilled than the small blockers. Whereas small blockers were better programmers, and promoted safe coding practices not to risk disruptive hard forks, replay risk, etc in addition to their focus on network consensus of both validating and mining nodes before changes could be enforced. Both sides played by the rules of engagement and dirty as well. Bitmain was not a paragon of ethics and secretly had an IP that was called ASIC boost, that allowed their miners to hash more efficiently and helped them have a profitability advantage. Segwit was a smaller blocker BIP soft fork initiative, that effectively increased the blocksize to 2Mb without increasing the 1Mb limit (don't ask me how). However, Segwit would neutralized ASIC boost's ability to use less electricity. The code base is maintained on Github by a few smaller blockers (1 main person at the time of the books writing), and there were several Reddit, Bitcoin talk, email lists, that discussed these issues (some maintained by small and others by big blockers). Ultimately, the smaller blockers won, but this outcome wasn't because they were any better at strategy and programming skill. They made some gambles that could have turned on them very badly if the big blockers saw these opportunities, if the big blockers had better computer science skills, if ethereum had not undergone a controversial hard fork that was very disruptive to recover stolen ETH by a hacker at around the same time this debate was occurring. Long and short of it, I think BTC will always be faced with politically/economically self-interest entities that want to centralize it in additional to the technical/developer risks of keeping BTC functioning on a decentralized manner. A complete history of Bitcoin's consensus forks - 2022 Update | BitMEX Blog Below is a problem that is yet to be fixed Long and short of this meandering post coupled with @wachtwoord's paper above, a few things come to mind. 1) payment systems don't accrue value like a store-of-value function, but big centralizing forces will always be a threat because they will capture the most value if BTC is mainly a payment system 2) how decentralized are the full nodes in the network out there? 3) who is going to update, maintaining and provide the necessary solutions for bugs, soft forks as this technology progresses through time and scale?