I was looking at Lemonade’s 10-K the other day and noticed something startling and wondered how I had never noticed it before.
Lemonade uses inappropriate accounting for its reinsurance which has the effect of inflating its revenue.
Importantly, there is no impact on net income, so this effect is purely cosmetic, but, don’t forget Lemonade doesn’t want you looking at its net income (recall, they use phony metrics like adjusted gross profit).
They do care about revenue as, for a time, the market valued them on a multiple of revenue. While I can’t prove the motivation for the accounting treatment, I suspect it was about boosting revenue to support valuation.
So what exactly did they do?
Accounting For Quota Shares
As most of you reading will know, when a company buys quota share reinsurance, the accounting treatment is to reduce premiums, losses, and expenses commensurately so that combined ratios on a gross and net basis are comparable.
In other words, if the ceding commission paid equals the cedant’s gross expense ratio, then the gross and net CRs will be the same.
For a simple example, if I have 100 of premium, 70 of loss and 25 expense, my gross CR is 95%.
If I buy a 20% quote share with a 25% ceding commission, I reduce premium by 20, losses by 14, and expenses by 5. This results in NPW of 80, net losses of 56, and net expenses of 20. My CR is still 95%.
For accounting presentation, this means I recognize the 14 of reduced losses as a benefit to my claims expense and the 5 of reduced expenses as contra expenses.
What did Lemonade do? They recognized the ceding commission as revenue, rather than a contra expense!
Ceding Commission Revenue
Now, some of you may be thinking, that can’t be right. Maybe I made a mistake or don’t understand their presentation. No, I assure you, I am accurately representing the facts.
It’s right there in the income statement – ceding commission income. You can see it is a meaningful component of revenue. It increased 2022 revenue by 1/3!
Wondering what counts as ceding commission revenue? Let’s go to the 10-K:
Ceding commission income is commission we receive based on the premium ceded to third-party reinsurers to reimburse us for acquisition and underwriting expenses.
We earn commissions on reinsurance premium ceded in a manner consistent with the recognition of the earned premium on the underlying insurance policies, on a pro-rata basis over the terms of the policies reinsured.
The portion of ceding commission income which represents reimbursement of successful acquisition costs related to the underlying policies is recorded as an offset to other insurance expense.
Lemonade 10-K, p.76
So it’s all of the cede other than what is attributable to customer acquisition (which they properly treat as contra expense). There is not enough information to fully break out how much of their cede they call revenue vs. contra expense, but clearly a significant amount counts as revenue.
They ceded $333M in ’22 and $389M in ’23. That means ceding commission revenue equaled 19% of ceded in ’22 and 18% in ’23.
Comparing Approaches
What happens to your ratios if you call ceding commission revenue instead of contra expense? Let’s take a look.
| 20% QS at 25% cede | Gross GAAP | Net GAAP | “Lemonade Style” Net GAAP |
| Premium | 100 | 80 | 80 |
| Revenue | 105 | 85 | 90 |
| Losses (LR) | 70 (70%) | 56 (70%) | 56 (70%) |
| Expenses (ER) | 25 (25%) | 20 (25%) | 25 (31%) |
| Underwriting Profit (CR) | 5 (95%) | 4 (95%) | -1 (101%) |
FYI, the difference between premium and revenue in the base case is NII and other ancillary income. The additional $5 of revenue in the Lemonade column is the ceding commission.
Note, the ceding commission is separate from earned premium, so it has no effect on net loss ratio. However, it obviously screws up the ER as you are comparing gross expenses to net premiums.
This increases the CR 6% in this example. You might object that Lemonade wouldn’t want to pursue accounting treatment that makes it CR look worse.
Who Needs A Combined Ratio?
Au, contraire! Lemonade does not report a combined ratio! I am not making this up.
Search their entire 10-K, earnings releases, investor day materials, you name it. You will never see the words “combined ratio“.
Have you ever heard of an insurer that doesn’t think combined ratio is an important metric? What kind of blasphemy is this? How do you expect people to take you seriously when you ignore the most important metric in insurance???
But, if your goal is to increase revenue, and if the tradeoff is making the CR worse, then I guess you pretend the CR doesn’t exist.
But What Is The Combined Ratio?
Fortunately, we have most of what we need to calculate Lemonade’s gross and net CR on our own.
| Gross ’23 | Net ’23 | Gross ’22 | Net ’22 | |
| Premium | 672 | 315 | 490 | 172 |
| Losses (LR) | 571 (85%) | 280 (89%) | 441 (90%) | 167 (97%) |
| Expenses (ER) | 379 (56%) | 309 (98%) | 384 (78%) | 320 (186%) |
| Underwriting Profit (CR) | -278 (141%) | -274 (187%) | -335 (168%) | -315 (283%) |
Even on a gross basis, Lemonade’s CR is well over 100%! As you can see, it is far worse on a net basis.
Note, these calculations properly take the ceding commission out of revenue and allocate it against expense. If I kept the Lemonade accounting treatment, those net CRs would be even higher!
When your net CR is around 200%, no wonder you don’t want to self report it! Better to goose your revenues instead, I guess.
Unprecedented?
I can’t recall, in all my years, ever seeing an insurer treat ceding commission as revenue instead of contra expense. It’s absurd to prioritize revenue over combined ratio, and it’s completely counterintuitive to the substance of the transaction.
If anyone has ever seen another insurer (an established one, not a recent startup) use this accounting, I’d love to see it. For those wondering, E&Y is Lemonade’s auditor, so if your company uses E&Y, maybe they can explain it to you.
The only case I can come up with where treating a cede as revenue would make sense is if you were a non-risk bearing MGA. In that case, you are passing along 100% of the risk, so there is no point in reporting premiums or losses.
Even then, I can argue that if you are incurring underwriting expenses to assess risk that meets the reinsurer’s guidelines or paying a retail broker commission to source risk, then the payment from the reinsurer should net those expenses out rather than be booked as revenue.
Overrides
So there’s one other issue I left out so far, which is the ceding commission doesn’t always equal the insurer’s actual expenses.
If I assume a 29% cede, aka a 4% override, then, using the example from above, net ER drops from 25% to 24%.
| 20% QS | Gross GAAP | Net GAAP @25% cede | Net GAAP @29% cede |
| Premium | 100 | 80 | 80 |
| Losses (LR) | 70 (70%) | 56 (70%) | 56 (70%) |
| Expenses (ER) | 25 (25%) | 20 (25%) | 19.2 (24%) |
| Underwriting Profit (CR) | 5 (95%) | 4 (95%) | 4.8 (94%) |
If one twists one’s self into knots, one could argue the extra 0.8 from the override is revenue and suggest they still have a 95% CR. This goes against all convention, but I could at least entertain a debate about it.
So, might this be what Lemonade is up to? Anecdotally, I’ve been told that’s how they explain this. They tell investors that the ceding commission revenue is an override.
But we know that doesn’t square with reality. Their gross ER is 56%! Yet, they are only getting an 18% cede (ex- the portion attributed to acquisition expense). This is not an override!
Even if I strip out all marketing and all tech expenses, they still have a 27% gross ER. The ceding commission does not cover their expense structure.
Reinsurers Aren’t Charities
Which makes intuitive sense. We know Lemonade’s ER is far too high and a reinsurer isn’t going to reimburse them for all the excessive spending.
Also, Lemonade’s LRs have run hot and a reinsurer can’t make a profit if they are giving Lemonade an override. Common sense says if the LR is in the 80s or 90s, then the ceding commission should be like 10% if the reinsurer is going to make a profit.
Thus, even the 18% seems high and suggests reinsurers are betting on future LR improvement. But it is unlikely the commission for the marketing expense is more than a token amount.
So, we can all hopefully agree Lemonade is not receiving an override and there is no rationalization to book any of this ceding commission as revenue.
In fact, if they really believe excess ceding commissions are revenue, then they should surely view deficient cedes as a subtraction from revenue! In that case, they should book a larger contra expense commensurate with their gross ER and any shortfall versus the cede they actually receive should lower revenue.
Investors Should Demand Better Disclosure
If you’re investing in Lemonade (or a sell side analyst covering it), you really should demand they conform to industry convention and restate their revenue and expenses.
You should also demand they start reporting their combined ratio.
It would also be helpful if they would disclose what the actual cede is they are receiving and be able to explain why it is so much less than their actual expenses.
It’s one thing to try to persuade investors to focus on your non-GAAP, largely irrelevant profitability metrics, but it’s completely unacceptable to not also provide basic financial metrics like expense ratios and combined ratios and to use very liberal revenue recognition practices that are completely out of line with peers.
If Lemonade wants to improve their credibility and move on from being the industry bad boy, one way to do so is to make a good faith effort to make themselves comparable to the industry and stop pretending they’re not an insurance company.
Another great piece, Ian.
Great piece as ever.
Regarding your unprecedented accounting point, State National used to report this way for their “Program Services” (fronting) business. But in their fronting model they retained 0% of the risk, i.e. 100% was reinsured, so their loss ratio on the business they wrote was zero. Their ceding commission equates to a fronting fee. See their last 10k here before they were acquired by Markel: https://www.sec.gov/Archives/edgar/data/1610793/000155837017001763/snc-20161231x10k.htm
So while the accounting was the same, State National is a fundamentally different business model from Lemonade, as they do not take any the underwriters.
Yes, they were who I had in mind when I was thinking of a proper usage for this treatment. I didn’t specifically know they used it, but I would have guessed someone with that model would likely have done so.
Thanks for the share on that. It’s interesting that E&Y was State Nat’s auditor too, so it looks like Lemonade is copying that treatment. Don’t know if they shopped for an auditor who would let them pretend they were a front or if E&Y sold it to them and they went along, but as you say, it’s a very different business model.
Ian – I have not done much research on them, but this would appear to be the company as an MGA as opposed a carrier.
The lack of a combined ratio would also be more consistent with a MGA construct.
Very interesting article. I am a former EY audit partner. Accounting rules are often prescriptive regarding revenue and expense recognition (i.e. things that impact net income) but surprisingly not prescriptive regarding geography. In other words, as you note this is a classification issue that has no impact on net income. So, my guess is that EY did not love this but could not say it is wrong. As long as it is clearly disclosed, which it is, then they are powerless to alter their opinion. Combined ratio is a non-GAAP measure so no issue with excluding it. I’ve seen lots of games with classification of expenses – such as inflating corporate expenses (cuz not in combined ratio), allocating some internal overheads to investment expenses (netted in investment income), ditto for burying some expenses in claims to reduce G&A, including internal UW costs in DAC and I’ve seen lots of games with various types of ceding commissions. But I have to say this accounting by Lemonade is a new one for me. No doubt motivated by the reasons you suggest.
Thanks, that’s great perspective. Totally agree some gaming of expenses happens at many big insurers around keeping expenses out of G&A and allocating them to places like corporate as you say.
Lemonade isn’t the first, and won’t be the last, to try to achieve a preferrable cosmetic outcome, but as we agree, this does seem to go further than usual. I thought I’ve seen pretty much everything in a 10-K but this one definitely made my head spin a bit!
Could the entire cede be an override because they’ve effectively offset their acquisition expenses with “synthetic agents”? That effectively swaps out opex for capex, so the cede is recognized entirely as revenue.
They’ve been doing this since 2020 so long before the synthetic agents.
Also, the part they recognize as rev is related to the G&A, not the CAC. As I showed, their G&A is well above the cede they receive, so there’s no realistic way to represent this as an override.
Typical MGA accounting (although there is some variability ) has the ceding commission booked as gross revenue and then net revenue excludes retail commission and other variable policy related costs. Profitability is most often focused on EBITDA as a percent of net revenue but sometimes people use gross.
If you treated all underwriting related expense as “above the line” as well then EBITDA would be very close to net revenue and wouldn’t really be a good measure of the MGAs operating profitability.
Yes, but Lemonade isn’t an MGA! That’s the main issue. They are accounting as if they are one, even while showing elsewhere in the financials they are a full stack insurer.
Also, the equivalent of commission in this case would be their CAC which they *do* treat as a contra expense, rather than as net revenue. So if you were to truly follow the spirit of the MGA approach, you would subtract most of the CAC from the cede…which would leave you with NEGATIVE net revenue! 🙂
PS: EBITDA isn’t GAAP. You would arrive at the same EBITDA regardless of how you present the income statement. EBITDA margin would change, but that perhaps show the fallibility in looking at EBITDA margins, if they are sensitive to the choice of accounting practice for revenue, rather than the underlying economics.
I was just referring to the comment you made about perhaps MGAs should treat UW as a contra expense – not related to Lemonade. I agree with you on EBITDA, and I think the reason that is such a common metric for MGAs is because they tend to be viewed through the eyes of an acquirer (PE or strategic, who value MGAs on a Multiple of EBITDA. You can definitely play with the margin number based on how you treat gross vs. net revenue.
If you recall back to the early days of Lemonade (pre-IPO) they did cede substantially all of their premium, around 90% if I recall correctly. The model most closely resembled the ill-fated Mutual Risk or early Tower Group.
Either just before they went public or just after they switched to a risk bearing model with a view to hyper-growth and a revenue based valuation focus (it was around here where they started using adjusted, adjusted earnings as well.
I would suspect that in the early days the accounting was reasonably appropriate and thats when EY signed off on it and somewhere along the way EY should have dug in their heels and said ceding commissions are material enough to force an accounting change – naturally managements hate restatements like that and no doubt that’s why it was resisted.
Good piece and just one more reason to dislike LMND.