Skip to main content

If You Tax Derivatives, Only Derivatives Will Pay Taxes

  • Author:
  • Updated:

In general when something is headlined "A Sensible Change in Taxing Derivatives," or "A Sensible Anything," that's a good sign that it's not; things that are sensible don't have to advertise. Also: ooh derivatives are evil ooh, so the odds are even worse. But this particular sensible change - a Victor Fleischer DealBook column about a Republican House proposal to tax derivatives on a mark-to-market rather than "open transaction" basis - is more sensible than you might initially expect; it's mostly plausible and inoffensive and non-pitchforky.1 (The idea is straightforward: if you own a derivative, and it went up in value this year, you pay taxes on the increase in value. Unlike with, say, stock, where you only pay taxes on the increase in value when you sell the stock.)

One way to tell it's not too bad is that various reports suggest that the Wall Street reaction so far has been "meh?" or "huh?"; this is presumably in part because it's not clear how real this is and in part because it's not clear how bad it'd be if it was real. Wall Street, in the sense of derivatives dealers, already pay taxes on their derivatives inventory on a mark-to-market basis, so the dealers' dog in this fight is not their own taxes but rather the marketability of various products, from boring ETNs to lovely variable share forwards, to customers focused on tax efficiency.

Nonetheless! There are two ways to think of derivatives. One is they are a specific class of evil things, often involving acronyms, designed to let banksters get up to dirty tricks. This line of thinking goes along with words like "complex" and "opaque" - "derivatives are complex instruments ..." etc.

The other way to think of the term is as a catch-all for any sort of contract whose value is determined in part by something in the world. By that measure most things are derivatives. Stocks are famously call options on companies. Many methods of financing financial instruments can be viewed as either loans - wholesome, wholesome loans - or derivatives - evil, evil derivatives.2 "Derivatives" is a useful organizational term - you can separate derivatives traders from traders of cash instruments - but doesn't particularly partition the world; derivatives traders and cash traders (stock!) and M&A bankers (contingent value rights!) all occasionally deal in things that you could, if you like, call derivatives.

This proposal sort of aims at the narrow - evil derivatives with dirty-tricks purposes - but defines with the general:

(a) IN GENERAL.—For purposes of this part, except as otherwise provided in this section, the term ‘derivative’ means —
(1) any evidence of an interest in (or any derivative financial instrument with respect to)—
(A) any share of stock in a corporation,
(B) any partnership or beneficial ownership interest in a partnership or trust,
(C) any note, bond, debenture, or other evidence of indebtedness,
(D) except as provided in subsection (e), any real property,
(E) any commodity which is actively traded (within the meaning of section 1092(d)(1)), or
(F) any currency ...

So a derivative includes "any evidence of an interest in any share of stock in a corporation." You know what's "evidence of an interest in any share of stock in a corporation"? Stock! Not just in the tautological way, either: nobody owns stock, they just own interests in their brokers' interests in DTCC's interest in stock. "Oh I own AAPL shares," you say, but you don't; you own like a second derivative on Apple shares. A delta-one derivative but still.

I assume that's a hypertechnicality and the rules wouldn't actually be interpreted that way,3 but it's an inauspicious start. What else is a derivative? Arguably a floating-rate bond is a derivative, since it is a contract whose payoff varies with respect to some index (Libor, whathaveyou). The proposed law would apply to "any embedded derivative component" of a bond, which would naïvely appear to apply to floating-ratiness.3.5 There is however an exception (section 486(d)(2)(B)(iii)) for variable rate debt instruments, so in fact your floating-rate debt will not be marked to market.4

Here I built you an interest-rate swap:

  • I issue and sell to you a floating-rate bond;
  • You issue and sell to me a fixed-rate bond;
  • The end.

Neither side is taxed on a mark-to-market basis: both are just bonds. The cash flows replicate a swap. (No?) Now you have a swap that isn't taxed on a mark-to-market basis. You can call it a "swap" and it's mark-to-market, or you can call it "two bonds" and it's not.

I poke these holes not with any particularly strong intent of sinking the proposal. You gotta tax something, and this seems plausibly designed to reduce the usage of certain transactions that are, as the euphemism goes, entered into mainly for their tax efficiency.5 It is - it strikes me as a weird view of the world: you can imaginably have a view of the world that says "financial instruments should be taxed on a mark-to-market basis," or a view of the world that says "financial instruments should be taxed on a realization basis," but taxing some via mark-to-market and some via realization seems both theoretically dubious and open to gaming.6 Fleischer writes:

Suppose for example that you approach an investment bank to purchase an option to acquire 1,000 shares of Facebook at $30 a share, just above today’s market price. Assume the option is long-dated and expires in five years. By the end of the first year, assume the value of Facebook has increased to $40 a share.

Under current law, the option contract is treated as an open transaction, and the option holder would not pay tax on the appreciation in value until the option is sold or, if exercised, until the underlying Facebook stock is sold.

But Mr. Camp’s proposal would require the option to be valued at the end of each year based on what the profit would be had the option been sold and repurchased it at the end of each year. If the option declined in value, you would have the benefit of a tax loss.

But if you just bought Facebook stock, you wouldn't have to pay tax on the increasing value (or get a loss on the decreasing value) until you sold it. Why does the difference make sense?7 And if it doesn't make sense - if there's no economic argument for putting one thing on the taxable side of the line and the other on the non-taxable side - then that probably means that there's a way around it.

Jeremy Stein gave a great speech today about overheating credit markets, financial regulation, and monetary policy; one of its many delights is the epistemic modesty about what can be measured and regulated. For instance:

This short-term financing share is difficult to measure comprehensively, but exhibit 5 presents one graph that gives some comfort. The graph shows dealer financing of corporate debt securities, much of which is done via short-term repurchase agreements (repos). This financing rose rapidly in the years prior to the crisis, then fell sharply, and remains well below its pre-crisis levels today. ...

Nevertheless, I want to urge caution here and, again, stress how hard it is to capture everything we'd like. As I said, ideally we would total all of the ways in which a given asset class is financed with short-term claims. Repos constitute one example, but there are others. ... One is naturally inclined to look at data on short-term debt like repo, given its prominence in the recent crisis. But precisely because it is being more closely monitored, there is the risk that next time around, the short-term claims may take another form.

This is part of a larger argument - that this epistemic modesty about regulation and monitoring of specific trades should lead to greater reliance on monetary policy as a regulatory lever - that is not quite relevant to our purposes. But that doesn't make it less right. When you focus on "derivatives," you cut down on the tax structuring done via "derivatives." But you just give tax structurers something else to structure around. You might make their job more challenging, but that's part of what makes it fun.

A Sensible Change in Taxing Derivatives [DealBook / Victor Fleischer]
Comprehensive Tax Reform [House Ways & Means]

1.It's not just derivatives, either; there are assorted other clean-up tax things. One day I'll bore you with my thoughts about what the proposal calls "the phantom taxable income problem associated with many debt restructurings" in which issuers face COD income for the change in fair value of restructured debt, even if principal isn't forgiven. The proposal claims "this problem has prolonged and intensified the past several economic downturns, including the recent financial crisis," and I'm willing to buy that some companies could have gotten bondholders to agree to restructurings that kept them afloat, but were deterred by the fact that they'd have taxable COD income. But: I suspect there is a lot of mischief that can be done here.

2.A put can be replicated with a margin loan. A call can be replicated with stock and a margin loan. That is all ye know on earth, and all ye need to know, for certain purposes.

3.What if they were? There are plenty of people who think that mark-to-market gains in all financial instruments (or, all assets) should be taxable each year; here is one prominent statement of the theory. It's a little extreme? In any case, if that was the intent, it'd be weird to bill it as "mark-to-market taxation of derivatives." It'd be "you pay taxes every year that your stocks go up, even if you don't sell them." It'd be kind of a radical change I think? That guy I just linked to, who advocated the change, said "Our tax system is based on the concept of 'realization.'" So if you got rid of that you'd have to rebuild from the ground up.

3.5.A reader points out a neat flaw here: call options in bonds would appear to be embedded derivatives. Most high-yield bonds, lots of munis, some investment-grade, etc. etc. have embedded calls where the company can take out the bonds at some fixed price or spread prior to maturity. These calls get more valuable as the bond gets more valuable. But of course the bond is not marked to market. What this means is that if you buy a HY 8NC4 bond for 100 in June, and it's worth 110 in December, then you've made, say, 12 points on the bond appreciating and lost (2) points on the call option becoming more valuable. (I.e. if it was an 8NCL bond it'd be worth 112. Or whatever. Numbers not important.) So you have an actual economic gain, but you can take a tax loss of 2 points. On the other hand if the bond went down to 90 you'd owe taxes on a 2-point (say) gain on the call option. That is ... very counterintuitive.

4.Best not to think too hard about this. To some extent, in theory, floating-ratiness shouldn't cause mark-to-market moves - holding credit constant, floating rate bonds should always be worth par(-ish) since they're always at market(-ish) rates. Fixed-rate bonds should have mark-to-market moves. But, whatever. You used the wrong index, or you have an inverse floater, or whatever.

5.I am looking at you lovely SunTrust trade. But mostly actually individuals doing it to avoid taxes til they die; Fleischer leads off with one example and there are tons.

6.Two clean-up points here. One, you can have a "tax everything on a mark-to-market basis" system: that's more or less what securities dealers have, and some people want it more broadly - see note 2. Two, we already have a system where some financial instruments - mainly listed futures - are taxed on a mark-to-market basis. And sometimes people use those, other times they use things that aren't taxed that way, and the choice is probably driven in part by tax efficiency.

7.What if Facebook issued warrants? That strikes me as a delightful rabbithole.