OK, disclaimer first. I have no interest in GE, either financially, reputationally, or in any other way. I have no axes to grind or allies to appease. I am writing solely because I think GE is being treated incredibly unfairly by the so-called “whistleblower”.

I understand Harry Markopolos was right on Madoff. I understand GE has, at times, been far from forthright about the problems at GE Capital, including long term care (LTC). I certainly wouldn’t be surprised if they still have to add to LTC reserves in the future (though not near the magnitude described). I have no problem with a short seller making his case publicly, if it is well researched and fair. I do have a problem when people intentionally mislead to try to scare people out of a stock. That is what is taking place here.

Questionable Motives

While Markopolos claims he is a whistleblower, buried in the middle (where it’s least likely to be seen) of the Wall Street Journal story is a disclosure that Markopolos is working for a hedge fund.

Mr. Markopolos said he and his colleagues are working with an undisclosed hedge fund, which is betting GE’s share price will decline. Mr. Markopolos’s group gave the investor access to the research prior to publication and will receive a portion of any trading proceeds. He declined to identify the hedge fund

Wall Street Journal

This should have been disclosed at the very beginning of the story in ALL CAPS and boldface type! This is not some justice-seeking citizen trying to help fight fraud. This is a standard short seller hit piece. Again, I have no problem shorting a stock where someone sees problems.

The issue here is the investor is hiding behind a front and letting someone else take all the heat. It also brings into question the independence and integrity of Markopolos’s work in the same way that people question environmental studies funded by oil companies.

The report cites that Markopolos is qualified to do this work because he is a Certified Fraud Examiner (CFE). I see on that association’s website there is a code of ethics. Accepting trading profits from a hedge fund using your research arguably violates several of the ethics standards.

Who is the Pot & Who is the Kettle?

The great irony in this “whistleblower” claim is Markopolos is guilty of the same behaviors he accuses GE of. Repeatedly, he accuses GE of deceit and manipulation. Yet, these are exactly the tools he uses in his report! He consistently overhypes accusations (GE is Madoff, GE is Enron) and takes things out of context (far too many examples to cite given this is an already lengthy writeup) to create impressions of problems that aren’t there.

I can only come to one of two conclusions: he either knows nothing about insurance and is making innocent mistakes, or, he knows more than enough to realize he is making false claims and is doing so intentionally in the hopes of pulling the wool over the eyes of readers who are less informed about insurance than he is.

Again, I am not vouching for the accuracy of GE’s LTC reserves. I have been skeptical back to the time of the Genworth spinoff. However, the time to write this report was two years ago when it was a significant issues, rather than piling on after it’s been largely addressed.

The Allegations

Markopolos claims that GE needs to add $29.0B to its existing $19.9B of GAAP reserves (slide 23). My conclusion is $19.5B of this is pure smoke and mirrors. The other $9.5B I would characterize as largely inflated, but perhaps having some merit.

GAAP vs. Stat

Let’s start with the GAAP-only issue first. Markopolos alleges that GE needs to add $10.5B to GAAP reserves so that they equal stat. First, there is absolutely nothing that says GAAP & stat reserves need to be equal or even close to each other. That’s a red herring.

Second, he says the coming changes to life insurance GAAP in 2021 (actually 2022 now, but who’s counting!) will force GE to eliminate this gap. No, it doesn’t. GAAP reserves will clearly go up but it has very little to do with stat. It is because GAAP will include more market based assumptions. That is not what stat does. It may be the reserves will be worse under GAAP than stat going forward (because GAAP will be more punitive for low interest rates)!

Now, to get to the heart of the issue. Markopolos alleges that because competitors have LTC reserves that are similar for GAAP & stat that GE must be hiding something from its GAAP results. On the contrary, they disclose it very clearly. You just have to pay attention. On page five of GE’s LTC investor presentation earlier this year, there is a footnote. That footnote shows a permitted practice from their regulator added $9B to statutory reserves. The $1.5B remaining difference is now in line with peers. Magic!

Now, that $9B will eventually makes its way into GAAP, but it is not from the “Enron playbook”. It’s plainly disclosed. One would think that someone who brags “most investors are not trained as Certified Fraud Examiners (CFE’s) and have no idea of what forensic accounting analysis entails” (page 6 of the memo) would have the skill to read the footnotes. But I’m not a trained CFE, so what do I know? Well, I know, for starters, we’ve eliminated $10.5B of the “fraud”.

JUDGMENT: $0 of the stat/gaap difference is problematic.


Reserves vs. Peers

Next, Markopolos moves onto stat and looks at the reserve disclosures of GE vs. some peers. Note, the emphasis on some. He cherry picks two and ignores the rest. More on this below.

His claim is that GE holds less reserves for older policies than Prudential and Unum. Now, this is a notoriously hard exercise to do. Stat disclosures in and of themselves are not up to the task. However, for now, let’s take them at face value.

Markopolos argues that GE’s ERAC block is under-reserved because it has $78,500 in reserves/life on active claims (more on this nuance below) and PRU has $113,500/life on its oldest active block. Multiplying this difference over the number of active lives yields a $9.5B “reserve deficiency”. He suggests this is a fair comparison because this is Pru’s oldest block and thus most similar to GE’s.

Like all good deceits, there are some true statements and compelling arguments above. There are also (purposely?) omitted points that would make the argument much less persuasive.

Yes, all else equal, it is fair to compare similarly aged blocks. However, why did he only focus on Pru and, to a lesser extent, Unum. Where does everyone else rank? Am I really to believe he didn’t look at all the other carriers? Or were the results counter to his thesis so they were conveniently ignored? If the latter, that’s pretty unethical in my book.

A better methodology would have been to include all older age blocks and take an average to compare to GE. I’d also note the Pru block he looked at only has 20,000 policyholders. Is that really significant enough to make a sweeping judgment on GE’s block which is more than 10X larger?

Let me return now to the focus on active lives. For the uninitiated, active lives are those NOT on claim. People who are making claims on their policy are called disabled lives. By definition, active life reserves are larger for older populations. This is common sense. There have been more premiums collected and set aside for their future claims. Thus, by picking Pru’s oldest block as a comp, he cherry picked the one that would be most adverse.

We also don’t know the composition of the two blocks as he did not show the attained age for the Pru block (he does show the age for their entire individual block, but not for the oldest block). It is possible the Pru block has more reserves because it has older ages. It is also possible that more of GE’s block has graduated to disabled claims leaving a healthier group of active lives behind than in Pru’s pool.

The point is there are a lot of unknowns to be leaping to broad conclusions. There are certainly too many unknowns to call any seeming differences “fraud”. That is flat out irresponsible.

JUDGMENT: Some of the $9.5B alleged reserve deficiency may be real, but it is too hard to tell. The magnitude is likely much smaller. Let’s be generous and call this inconclusive.

UPDATE: After getting access to the data, my hunch was right that Pru and Unum’s selected blocks were the worst in the entire industry. It was 100% cherry picking! I calculated the average reserve for 49 “old” blocks and the reserve/life was $39,700, so almost 1/2 of GE’s!!!


The Premium Deficiency

Next, Markopolos alleges a $3.6B deficiency because GE’s pricing on its policies is too low. This is based on, again, a comparison to the Pru block. He suggests that because GE only collects $1,133 on average vs. $2,891 for a Pru customer that we should PV this difference over time and thus GE needs $3.6B more of premium.

This is double counting at its finest! It may be true that GE is underpriced (or not, see below) but that has nothing to do with reserve adequacy. The necessary reserves are the necessary reserves. How much you charge for them has nothing to do with it. It will effect your profitability but not your reserve adequacy. In other words, when GE did its study in 2017, the amount of premium it expected to receive was a given, not as assumption. There is nothing to adjust. The only thing to adjust is claims.

The only reason this would change is if they could raise prices. However, this would lower the reserve deficiency, rather than reduce the new $3.6B. If they could raise pricing to Pru’s level, then the $3.6B mentioned here would still be $0 but the $9.5B would decline to $5.9B.

So, we can throw this whole $3.6B out the window! But, before we do, it is enlightening in another way. Perhaps the reason GE has lower reserves/policy than Pru are because it has lower premium/policy? In other words, maybe there is a mix issue of some sort that would cause GE to have customers who pay less but can claim less. If so, this would be one more reason to cast doubt on the $9.5B above.

JUDGMENT: $0 of this $3.6B is legitimate.


The Remainder

The last $5.4B is a real piece of work! With a straight face (well, I can’t see his face, only the numbers on the page, but permit me the expression), Markopolos throws in a last $5.4B gut punch.

Why? He throws a 15% “risk factor” on top of the entire reserve base (including the $13B he invented above = more double counting). Why? Because he felt like it. There is no substantiation provided! It reeks of wanting to put a scarier number out there. It’s appalling!

JUDGMENT: Most definitely $0.


The Final Tally

I have clearly dismissed $19.5B of the total. The other $9.5B may have some validity but likely at a smaller size. Let’s arbitrarily (since that is the theme of the day) take half of it and say the deficiency is between $0 and $4.75B. That is material, but certainly manageable.


UPDATE: See above for updated information. I am lowering the $4.75B to approximately $0!


This report is a propaganda tool for a short seller. If this is “forensic” accounting, then it is the equivalent of someone planting evidence to falsely incriminate. I would be embarrassed to publish a report of this quality under my name.

An honest report would present all the facts rather than cherry pick. It wouldn’t create distractions like screaming Enron and Madoff. It wouldn’t ascribe motives to GE’s behavior when there is no way to verify them. It wouldn’t present charts that have clearly misleading representations. It would admit that there is much that we can’t know as outsiders.

There is certainly a strong case to be made that GE was dishonest about its LTC reserves over the last 15 years. However, the time for that report was two years ago. This report is yesterday’s news dressed up to look new. It adds no value. Worse, it deceives the average reader.

3 thoughts on “GE “Whistleblower”: Who Is the Real Fraud?”

  1. Really the only real way to get at this would be objective actuarial reviews to assess underwriting quality, and review the performance quality of the insurance portfolio. What does concern me is the the quote about a loss ratio reported to be in the stratosphere.

    1. The Loss Ratio slide (p74) is also misleading. He compares the GE blocks to the total blocks of competitors, even though he suggests elsewhere that it should be compared to the oldest blocks of peers. Well, right on the next page (75), he shows you the LR of the worst PRU block is 185% and the worst Unum is 363%. Now, the GE blocks are high, but not out of line. The document is riddled with cherrypicking like this!

Comments are closed.