There is a rule, called Regulation FD, that says that U.S. public companies cannot selectively disclose material nonpublic information to some analysts or investors without disclosing it publicly. So a chief executive officer can’t meet with a big mutual fund, or a Wall Street analyst, and say “our earnings will be $1.75 per share this quarter,” unless her company has disclosed those earnings publicly.¹ It is a weird set of stylized facts that: 

1. Public-company executives meet constantly with their big shareholders;
2. The shareholders value these meetings, as you can tell because (a) they go to them and (b) they reward Wall Street banks for setting them up;
3. There is some evidence that shareholders who go to these meetings ask useful questions and outperform shareholders who don’t; and
4. Cases of Regulation FD enforcement are quite rare.

What do they talk about in these meetings, which investors find so valuable, if not material nonpublic information? The weather? Information that is immaterial, but that can be combined with other immaterial information to become material? In insider-trading lore this is called the “mosaic theory”; in philosophy I believe it is called the “sorites paradox.”

Anyway Reg FD was adopted in 2000 not so much to stop companies from telling their favored shareholders stuff that they didn’t tell everyone (though that too), but to stop them from telling Wall Street analysts stuff that they didn’t tell everyone. Before Reg FD it was sort of an accepted casual theory that the main channel for public companies to communicate with investors was through analysts. The analysts were in charge of explaining the company to the investing public, and it was important for the company to talk to the analysts so that they got the explanation right. If, for instance, the analysts all wrote reports saying “Amalgamated Widgets will make $2.15 per share this quarter on 20% gross margins,” and Amalgamated Widgets knew it was actually going to make $1.75 on 17% gross margins, Amalgamated Widgets would call up the analysts saying “you’ve got too high a gross margin in your model,” and the analysts would fix their models, and the quality of public information about Amalgamated Widgets would be better.

This was always a weird theory: Analyst reports are not exactly public (each analyst’s reports are generally provided only to clients of that analyst’s bank), and if a company wanted the public information about it to be better, it could always just release it publicly. In practice though it is awkward for a company to put out a press release saying, like, “we’re going to release earnings in two weeks but FYI gross margins will be a little lower than you think”; it is easy—or it was—for the company to call up a dozen sell-side analysts and say that. So companies made occasional formal public announcements to communicate big definite things, and they made more frequent informal private calls to analysts to communicate smaller less certain things. 

You can’t do that anymore, and that’s probably an improvement, but something was lost.

AT&T Inc. allegedly keeps it old-school though. Here’s a U.S. Securities and Exchange Commission enforcement action from Friday:

According to the SEC's complaint, AT&T learned in March 2016 that a steeper-than-expected decline in its first quarter smartphone sales would cause AT&T's revenue to fall short of analysts' estimates for the quarter. The complaint alleges that to avoid falling short of the consensus revenue estimate for the third consecutive quarter, AT&T Investor Relations executives Christopher Womack, Michael Black, and Kent Evans made private, one-on-one phone calls to analysts at approximately 20 separate firms.  On these calls, the AT&T executives allegedly disclosed AT&T's internal smartphone sales data and the impact of that data on internal revenue metrics, despite the fact that internal documents specifically informed Investor Relations personnel that AT&T's revenue and sales of smartphones were types of information generally considered “material” to AT&T investors, and therefore prohibited from selective disclosure under Regulation FD. The complaint further alleges that as a result of what they were told on these calls, the analysts substantially reduced their revenue forecasts, leading to the overall consensus revenue estimate falling to just below the level that AT&T ultimately reported to the public on April 26, 2016.

The complaint is worth reading to understand how public companies’ investor relations (IR) departments work. In the world of financial journalism, you will occasionally meet reporters and editors who argue that it is wrong to say “AT&T missed earnings estimates,” because—they argue—companies can’t miss estimates; if a company’s actual earnings differ from estimates, that means that the estimates missed, that the analysts got it wrong. AT&T does not experience life that way! AT&T’s goal in life, alleges the SEC, was to make sure it didn’t miss estimates:

Negative headlines generated by AT&T’s 4Q15 revenue miss in January 2016 caused consternation within AT&T’s IR department.

For example, the IR Director, who was located in Dallas, Texas, urged his team to take steps to avoid another revenue miss in the following quarter (i.e., 1Q16), writing in an email to Womack: “Chris, we have a tendency to focus on EPS and have recently missed the mark on consolidated revenue. We need to make sure our story gets consensus trued up for both EPS as well as revenue.” When Womack responded with an analysis of analysts’ revenue estimates for 2016 showing that the consensus revenue estimate for 1Q16 was higher than AT&T’s internal forecast at the time, the IR Director replied: “We will have to nip 1Q in the bud, otherwise we will be in the same spot we’ve been the last few quarters, i.e. missing revenue.”

AT&T did not experience its “revenue miss” as a failure by analysts to estimate revenue correctly. Nor did it exactly experience it as a failure by AT&T to earn enough revenue. It experienced it as a failure by AT&T to tell its story correctly, through analysts, so that the analysts’ revenue estimates would be just a hair below what AT&T actually reported.

So when it seemed likely that they would miss revenue estimates for the first quarter of 2016, they tried to “nip 1Q in the bud” by normal legal means, i.e. announcing in some public forum that revenue would be lower than expected:

After receiving these actual quarter-to-date results and learning that wireless equipment revenue and upgrade rates were on track to fall below analysts’ forecasts, AT&T determined to make a public disclosure to manage market expectations.

AT&T briefly considered issuing a Form 8-K to address, among other things, the accelerating downward trend in its upgrade rate and wireless equipment revenue. It elected instead, however, to have its CFO address the issue at an investor conference on March 9, 2016, at which the CFO was already scheduled to speak. 

But then they sort of chickened out of actually saying what they meant:

During the CFO’s remarks at the March 9 conference, which were webcast, the CFO referred back to his comments from AT&T’s 4Q15 earnings release regarding the decline in wireless equipment revenue and stated that he “would not be surprised” to see that trend continue. 

The CFO gave no quantitative guidance about AT&T’s wireless equipment revenue or any other metrics for 1Q16. Asked specifically by the conference host to do so, AT&T’s CFO stated that he could “only talk about up through the fourth quarter” and repeated the figures he relayed in January 2016 during AT&T’s 4Q15 earnings call.

Basically they wanted to say publicly “hey revenue is going to be even worse than you expect,” but they felt uncomfortable actually doing that. It feels scary, somehow, for a CFO to give specific forward guidance at a conference; that feels like something you should say officially, in writing, or not at all. On the other hand what you probably should not do is this:

Following the CFO’s remarks at the March 9 conference, AT&T’s IR Department developed a plan to contact individual analyst firms whose estimates were higher than AT&T’s projections. The purpose of this outreach was to get each analyst firm to lower its revenue estimate sufficiently to bring the resulting consensus estimate down to the level that AT&T expected to report.

AT&T and Womack, Evans, and Black understood that they needed the analyst firms to lower their revenue forecasts by a total dollar amount that, in the aggregate, was large enough to lower the consensus estimate to an amount AT&T could meet. 

And:
As Womack, Evans, and Black also understood, the IR Department’s outreach to analysts and its objectives were a top priority at the company. For example, after receiving a report on March 22, 2016, showing that analysts were still forecasting AT&T’s 1Q16 revenue to be over $1 billion higher than AT&T’s internal estimate, the CFO stopped by the office of the IR Director to make sure that his team was “working the analysts that still have equipment revenue too high.” The CFO was assured that it was the IR Department’s “top priority over the next few weeks.” 

The CFO just talked at a conference where he knew how bad revenue would be, and was asked about it, but refused to say! And then he stopped by the IR department to tell them how important it was for them to call analysts and explain that revenue would be low.

Anyway it really was a priority; the IR team “held private, one-on-one phone calls with approximately 20 sell-side analyst firms covering AT&T,” in which they allegedly “disclosed, among other things, AT&T’s internal upgrade rate and wireless equipment revenue data for the first quarter on these calls and otherwise communicated to the analysts, in sum and substance, that the analysts’ revenue estimates were above what AT&T was expecting to report and therefore needed to be reduced,” and “each of the twenty analysts issued revised research reports reducing their revenue estimates shortly after the calls.” And it worked, barely:

On April 25, 2016—the day before AT&T reported its 1Q16 earnings—the last of the approximately 20 analyst revenue reductions brought the consensus estimate just below what AT&T knew it would ultimately report.

AT&T’s CFO emailed the IR Director: “We may just beat revenue consensus.”

The IR Director replied: “I think we will :)”

The CFO forwarded this email to AT&T’s Chief Executive Officer and wrote: “These two updates may do it for us—we may beat revenue consensus—not by much but a beat nonetheless.”

The CEO replied, “Good.”

AT&T’s reported $40.535 billion in revenue for 1Q16, beating the final consensus revenue estimate by less than $100 million. 

From an efficient-markets perspective this is all bizarre. One day before AT&T reported its results, the revenue was what it was. Reporting a lot of revenue would be good; reporting less revenue would be less good. But one day before the results, the CFO and the IR director were not worried about the revenue; they were worried about the estimates. Getting analysts to lower their estimates—to say that AT&T would make less money—was good, because then, when AT&T reported how much money it made (again, a number that it already knew), that number would be higher than the consensus estimate, which would allow everyone to say that AT&T “beat revenue consensus.” ²  AT&T’s actual revenue, reported on April 26, would be higher than the average of analyst estimates on April 25, though not higher than the average of those estimates on April 24. It is a completely arbitrary bar, but it mattered to AT&T.

AT&T completely disagrees with all of this and plans to fight:

Tellingly, after spending four years investigating this matter, the SEC does not cite a single witness involved in any of these analyst calls who believes that material nonpublic information was conveyed to them.

The information discussed during these March and April 2016 conversations concerned the widely reported, industry-wide phase-out of subsidy programs for new smartphone purchases and the impact of this trend on smartphone upgrade rates and equipment revenue. Not surprisingly, without device subsidies, customers upgraded their smartphones less frequently, leading to a reduction in equipment revenue.

Not only did AT&T publicly disclose this trend on multiple occasions before the analyst calls in question, but AT&T also made clear that the declining phone sales had no material impact on its earnings. Analysts and the news media frequently wrote about this trend and investors understood that AT&T's core business was selling connectivity (i.e., wireless service plans), not devices, and that smartphone sales were immaterial to the company's earnings.

The evidence could not be clearer—and the lack of any market reaction to AT&T's first quarter 2016 results confirms—there was no disclosure of material nonpublic information and no violation of Regulation FD.

I mean, it’s possible that the lack of any market reaction was because AT&T guided expectations precisely to its actual revenue? 

How’s GameStop Doing?
Okay I give up: Why hasn’t GameStop Corp. sold a bunch of stock to the Reddit-based retail investors who are clamoring to buy it? We’ve been discussing this question since I first wrote about GameStop on Jan. 25, and last month we got what I thought was a definitive answer: During the big rally of late January, GameStop had too much nonpublic information about its fourth-quarter earnings to do an equity offering, but it had not buttoned up that information sufficiently to disclose it. GameStop’s fiscal year ended on Jan. 30, and by the late-January rally its executives knew a lot about how the quarter had gone. If they did a stock offering without disclosing that information, they’d get in trouble. But the information was still pretty preliminary, and if they did disclose it and it turned out to be wrong, they’d also get in trouble. Rushing to disclose it was too risky. So they did nothing.

Now it’s March! Pre-disclosing earnings in January would have been rushed and risky, but by now GameStop has had a lot of time to get the earnings right. It lost its chief financial officer two weeks ago, which is awkward, but still. On a normal schedule, GameStop will announce earnings in two weeks or so.³ It might have been a good idea for GameStop to spend the last, like, month and a half getting ready to do that a little early, so that it could peel off an at-the-market equity offering whenever it wanted to. 

The iron is hot, man:

GameStop Corp.’s latest winning streak has catapulted it to the highest in weeks as Chewy Inc. founder and activist investor Ryan Cohen continued to shake up operations at the video-game retailer, taking retail investors by storm.

Shares of the Grapevine, Texas-based company climbed as much as 19% to $231.74 at 9:36 a.m. in New York Tuesday, extending a winning streak for a fifth day. The gain follows Monday’s 41% climb after the company said Cohen would lead a new committee focused on its digital transformation.

Retail investors’ Reddit-driven frenzy has kept its shares afloat despite the broader market’s recent volatility and tech selloff. The stock has soared as much as 96% over five days. While the Nasdaq 100 fell 11% into a correction yesterday, it has rebounded by as much as 2.9% on Tuesday. Even so, GameStop remains far from its Jan. 28 intraday record of $483.

GameStop closed at $194.50 yesterday; it opened at $217.71 today; it got as high as $244 at 10:39 a.m. At that last price it is up 1,195% year-to-date and has a market capitalization of about $17 billion. At those prices, GameStop should go sell $1.5 billion of stock, travel back in time to December, buy up 100% of its stock and go private. I guess that wouldn’t work. It should sell $1.5 billion of stock, travel forward in time to May, and buy up 100% of its stock? It should sell warrants and puts? I don’t have a good answer, really, but it feels like these are the questions to ask. 

Here is GameStop’s press release announcing that Cohen will chair its Strategic Planning and Capital Allocation Committee “to identify initiatives that can further accelerate the Company’s transformation.” Presumably those initiatives will be in the vein of “do more stuff to pivot to online game retail faster,” but the committee does have “capital allocation” in its name. Why not do a billion-dollar at-the-money offering, get a huge mess of capital to allocate and then, I don’t know, buy an electric-spacecraft startup? GameStop arguably has access to capital at some of the cheapest terms in financial history; it seems weird to use that capital sparingly and on video-game retail. People are desperate to buy GameStop shares at any price! Take their money! Live a little!

I tell you what, whether it’s tomorrow or in two weeks, GameStop’s fourth-quarter earnings are going to be the financial event of the season. Will it announce a capital raise alongside earnings? Will the earnings be … good? Could it possibly matter? If GameStop misses estimates by a penny—Bloomberg tells me the consensus is about $1.43 of adjusted earnings per share on about $2.3 billion of revenue—will the whole thing vanish in a puff of smoke? “Sure I have diamond hands but they missed earnings so I’m out.” Somehow I doubt it.

Incentives
Here’s a good article about why corporate treasurers mostly aren’t putting their companies’ money into Bitcoin. One answer, which we have talked about before, is the bad accounting for Bitcoin. If a company buys Bitcoin, it books it as an intangible asset; if the price of Bitcoin goes down, the company writes down the asset and books a loss in earnings; if it goes up, the company does not book any gain. For a volatile asset like Bitcoin, this sort of asymmetric accounting treatment is very unattractive: It probably will go up a lot and also down a lot, which means that you will win some and lose some, but as far as your earnings go you will only lose some.

Another answer, though, is the asymmetric incentives that corporate treasurers face:

“The general consensus among treasurers is that very few of them are going to follow this trend initially,” said Naresh Aggarwal at the UK’s Association for Corporate Treasurers.

“As a treasurer, if I am right and the price doubles, the company may sell its holding and make a profit. Whilst the company may be worth more, it won’t be reflected in my compensation,” he added.

“But if the price falls, I am pretty confident I will be fired. Why bother putting my neck on the line?”

I am so used to the asymmetric incentives of the financial industry, where if you make a lot of money for the firm you get rich, and if you lose a lot of money for the firm you (1) get fired, (2) get to keep the riches you made previously, and (3) easily get another job because you have proved yourself as a risk-taker. These incentives, obviously, encourage risk-taking, which in some very broad sense is what the financial industry wants; investment banks and hedge funds are trying to identify and train skilled risk-takers. One can go too far with this, and over the last decade or so there has been a lot of conversation about dialing back these incentives, about bonus caps and clawbacks to reduce the asymmetry and discourage excessive risk-taking. 

But the incentives of the corporate treasury industry are the opposite. Your job is not to lose money. If you succeed in not losing money, you receive your modest, stable rewards. If you fail and lose money, you are fired. If you succeed beyond your wildest dreams and make a lot of money, (1) you do not get any extra rewards (because that was not your job) and (2) if your bosses are smart, they fire you anyway, because the fact that you made a lot of money means that you took a lot of risk, which was very much not your job.

The conclusion here is probably that corporate treasurers are never going to decide to buy Bitcoin, at least not until it becomes broadly normalized. How can it become normalized if treasurers never decide to buy it? Well, corporate treasurers are not necessarily the final decision-makers about the corporate treasury. Corporate chief executive officers are chosen and trained and incentivized to take risks; their pay does go up if the company is worth more. Some of them are really into Bitcoin. If Elon Musk tells his corporate treasurer to buy Bitcoin, Tesla Inc. is going to buy Bitcoin. “Ehh I just work here,” the treasurer will shrug; “they don’t pay me to take risks, and telling Elon not to do this is a risk I’m not willing to take.”

Extremely Pleasing Tax Case
The way corporate income tax works is pretty much that you add up the company’s revenue, you subtract its expenses, and you’re left with its net income. Then it pays taxes as a percentage of that net income. This is basically straightforward, though sometimes people fight over what should count as an expense. (In the U.S., interest on debt is tax-deductible, but dividends on stock are not, which arguably encourages excessive debt, etc.)

The way personal income tax works is not like that. You add up all your revenue, but you do not subtract your expenses, generally; if you are paid $100,000 a year and spend $93,000 on food and clothes and rent and stuff, your taxable income is $100,000, not $7,000. Well, that’s not quite right either; there are a bunch of expenses that are “deductible”—they reduce your taxable income—and sometimes people fight over what expenses should be deductible. (In the U.S. these fights tend to be over things like mortgage interest, state taxes, charitable contributions, etc.)

But broadly speaking, in the U.S., most of the expenses of life are not deductible, and people get taxed on their gross income, not the income they have left over after buying stuff. But what is income? Obviously if you have a job and get a salary, that’s income. If you buy a lottery ticket and win, your winnings are income. Most of the time, if you get money, that’s income. But if you buy some shoes and don’t like them and return them, and the shoe store gives you a $100 refund, surely that’s not income. Your shoe purchase was not tax-deductible; you couldn’t reduce your taxable income by $100 by buying shoes. So when you return the shoes and get your money back, that should not increase your taxable income by $100.

Similarly if you buy a product and send back the mail-in rebate coupon and get back $5, that's not income. Or if you buy a product on your credit card that gives you 5% cash back, that 5% cash back is not income; it’s just a reduction in the purchase price. There is a “longstanding Internal Revenue Service practice, which says credit-card rewards are usually nontaxable rebates. In other words, buying a pair of shoes for $100 and getting a 5% reward is really a $95 purchase, not $5 of income.” ⁴

What if the product that you buy is $6.4 million worth of grocery-store gift cards that you then exchange for money orders that you then use to pay off your credit card bill, allowing you to keep the 5% cash reward offered by the credit card company? Then:
1. You are amazing, my friend, truly, hats off to you.
2. But is it taxable income?

The person who did this is named Konstantin Anikeev, and he is an experimental physicist. The Wall Street Journal, which tells the story of his gift-card experiment, calls it “an inquiry far outside his field,” but I disagree; he built a perpetual motion machine, which has fascinated experimental physicists for centuries. He was able to “exploit the difference between unlimited 5% rewards and lower fees on gift cards and money orders” to just shuffle money around, risklessly, routinely, in a way that left him with more money than he started with:

His American Express card offered unlimited 5% rewards at grocery stores and pharmacies after he had spent $6,500. So Mr. Anikeev used his AmEx card to buy prepaid Visa gift cards at grocery stores, routinely stopping during his commute and purchasing the maximum allowed per day at a store. He often used the gift cards to buy money orders, then used the money orders to make deposits in his bank account, then used that money to pay his credit-card bill.

In a $500 transaction, the 5% rewards would yield $25—more than enough to cover gift-card fees of about $5 and the $1 fee on the money order.

He made about $310,000 doing this.⁵ As an exploit of credit card rewards it is fairly straightforward, though ambitious in its sheer scale. (An American Express spokesman “said the company uses a ‘combination of strategies’ to police the rules of rewards programs that don’t allow purchases of cash equivalents.” ⁶ )

As an inquiry into the nature of taxable income it is more profound, though. The IRS went after him, claiming that the $310,000 was income and he should have to pay taxes on it. He replied, no, it was just rebates on purchases. The IRS is sort of obviously right. Surely the only way to get $310,000 of rebates on purchases is by having more than $310,000 of taxable income. If you make $3.1 million at your job and buy $3.1 million of stuff and get 10% cash back then, sure, you got $310,000 of rebates, and those rebates are not income. If you make $100,000 in salary and cycle it frantically through gift cards until you earn $310,000 of rewards, it can’t really be the case that your taxable income was only $100,000. If that $310,000 was just a reduction in the price of goods that you purchased, where did you get all the money to do all that purchasing? Surely from income, right? ⁷

The IRS mostly lost in Tax Court though; here is the opinion. Basically the judge found that the perpetual-motion machine works and is non-taxable; if you buy gift cards with your credit card, use the gift cards to buy money orders, and use the money orders to pay your credit card bill, then it’s not taxable income. ⁸  Seems wrong philosophically, but it’s based on the IRS policy that credit card rewards are not income. The judge writes: 

This policy reflects the recognition that a taxpayer who avails himself or herself of a discount in acquiring goods and services has no accession to wealth. That taxpayer has retained more of his or her wealth than a taxpayer who pays full price for the same good or service, but that taxpayer has no additional income; he or she simply has reduced consumption. Although as Benjamin Franklin wisely observed, “[a] penny saved is two pence clear” (which became known more colloquially as “a penny saved is a penny earned”), the income tax law imposes a tax only on the future penny earned, not on the current penny saved.

That obviously is not what happened here—he did not reduce consumption—but if you have a general theory it will not always fit all the specific facts. Here the general theory is “rebates on consumption are not income.” You set up some extreme conditions, you run some experiments, and you see if the theory holds. Weirdly, it does.

Things Happen
Amazon-Backed Deliveroo Plans IPO, Testing London’s Appetite for Big Tech. Quantum-Computing Startup IonQ Plans Public Debut in $2 Billion SPAC Merger. Handling $1,109 of transactions “made the Mavericks ‘the LARGEST #DOGECOIN MERCHANT IN THE WORLD!’” How to Defeat a Boston Dynamics Robot in Mortal Combat. Trader Buys $36 Million of Copper and Gets Painted Rocks Instead. RIP Norton Juster. “Shoe Zone has named Terry Boot as its next finance boss after predecessor Peter Foot walked away from the role.”

Matt Levine is a Bloomberg Opinion columnist covering finance. He was an editor of Dealbreaker, an investment banker at Goldman Sachs, a mergers and acquisitions lawyer at Wachtell, Lipton, Rosen & Katz, and a clerk for the U.S. Court of Appeals for the 3rd Circuit.

1  Technically the rule is that the company has to disclose it “simultaneously, in the case of an intentional disclosure; and promptly, in the case of a non-intentional disclosure.” So if she’s going to meet a hedge fund and say “we made $1.75 this quarter,” the company needs to put out a press release simultaneously—more practically, *before* she meets the hedge fund—saying that. If she meets the hedge fund and accidentally blurts it out, the company can put out the press release as soon as she leaves the meeting. But best practice is to put it out before the meeting so there’s no doubt.

2  Byrne Hobart points out that AT&T’s alleged behavior is “healthier than the *old* way of beating estimates,” which was basically earnings manipulation and accounting fraud.“If AT&T really wants to come in a penny or two ahead of consensus, better to do that by browbeating consensus than by finding extra profits that don't really exist.”

3  The legal deadline for GameStop to file its 10-K with audited financial results for its last fiscal year is 75 days after the year ended, or mid-April. (GameStop is an “accelerated filer,” which means its 10-K is due 75 days after year-end; “large accelerated filers” have to file within 60 days, and that is the more normal filing time for big public companies. Earnings generally come before or alongside the 10-K.) Last year GameStop announced Q4 earnings on March 26, a bit less than two months after its fiscal-year end. GameStop doesn’t seem to have said when it will announce this year’s earnings, but Nasdaq expects earnings on March 25; Bloomberg’s ERN page says March 26.

4  To be clear, all this stuffdoesn’t matter for net-income-based corporate taxation. If a company has $500 of revenue and buys $100 of shoes and gets a$5 rebate, then its net income is $405, and it pays taxes on that, and it just doesn’t matter if the $5 increases “income” or reduces “expenses.” What makes personal taxation harder is that expenses are not generally deductible.

5  Of this, “$535 and $894 of the rewards accumulated in 2013 and 2014, respectively, were for the purchase of products other than Visa gift cards, debit cards, and money orders and thus are not subject to tax,” says the Tax Court. But $35,665 in 2013 and $276,381 in 2014 were from the disputed gift-card transactions.

6  Who is the “victim” here? Who paid him the $310,000? In a sense the answer is American Express, but it doesn’t seem to have cared. Anikeev’s “actions never offended American Express,” says the judge, and why should it? Presumably Amex charged grocery stores at least as much in interchange fees as it paid in rewards. Presumably the losers here were the grocery stores who sold $500 gift cards for $475 or something (net of Amex interchange fees) and then had them redeemed for $500 money orders.

7  No, fine, you can get it from savings—from previously taxed income—but that is clearly not what happened here either.

8  If you buy money orders directly with your credit card, the cash back on *those* purchases is taxable; Anikeev did some of that too.