POSNER, Circuit Judge.

Gary Bielfeldt (and his wife, but she is a party only by virtue of having filed a joint return with her husband), a large trader in U.S. Treasury notes and bonds, seeks to overturn a decision by the Tax Court denying him the right to offset immense trading losses that he incurred in the 1980s against all but $3,000 a year in ordinary income. He claims to be not a trader but a dealer and that the losses he incurred in the sale of the Treasury securities were losses connected with his dealer’s “stock in trade”; such losses, even when they result as his did from the sale of a capital asset, are treated as ordinary rather than capital losses and can therefore be fully offset against ordinary income. 26 U.S.C. § 1221(1). In contrast, capital losses, while they can be fully offset against capital gains, can be offset against ordinary income only up to $3,000 a year. 26 U.S.C. § 1211(b); Marrin v. IRS, 147 F.3d 147, 150-52 (2d Cir.1998). Although the amount is arbitrary, the rationale for limiting such offsets is not; it is to reduce *1037 taxpayers’ incentives to so structure then-capital transactions as to realize losses today and defer gains to the future. See Robert H. Scarborough, “Risk, Diversification, and the Design of Loss Limitations Under a Realization-Based Income Tax,” 48 Tax L.Rev. 677, 680-81 (1993); Alvin C. Warren, Jr., “The Deductibility by Individuals of Capital Losses Under the Federal Income Tax,” 40 U. Chi. L.Rev. 291, 310-14 (1973). If Bielfeldt’s characterization of his status is sound, he is entitled to some $85 million in refunds of his federal income tax.

The standard distinction between a dealer and a trader is that the dealer’s income is based on the service he provides in the chain of distribution of the goods he buys and resells, rather than on fluctuations in the market value of those goods, while the trader’s income is based not on any service he provides but rather on, precisely, fluctuations in the market value of the securities or other assets that he transacts in. Marrin v. IRS, supra, 147 F.3d at 151; United States v. Wood, 943 F.2d 1048, 1051-52 (9th Cir.1991); United States v. Diamond, 788 F.2d 1025, 1029 (4th Cir.1986). This is not to deny that a trader, whether he is a speculator, a hedger, or an arbitrageur, serves the financial system by tending through his activities to bring prices closer to underlying values, by supplying liquidity, and by satisfying different preferences with regard to risk; he is not a parasite, as the communists believed. But he is not paid for these services. His income from trading depends on changes in the market value of his securities between the time he acquired them and the time he sells them.

Although one thinks of a dealer’s inventory or stock in trade as made up of physical assets; it can be made up of securities instead. A stockbroker who owned shares that he sold to his customers at market price plus a commission would be a bona fide dealer. The example of a recognized dealer in securities that is closest to Bielfeldt’s self-description because it blurs the distinction between deriving income from providing a service in the purchase or sale of an asset and deriving income from changes in the market value of an asset is a floor specialist on one of the stock exchanges. The specialist maintains an inventory in a specified stock in order to maintain liquidity. If its price soars, indicating that demand is outrunning supply, he sells from his inventory to meet the additional demand, and if the price of the stock plunges, he buys in the open market in order to provide a market for the people who are trying to sell. See Bradford v. United States, 195 Ct.Cl. 500, 444 F.2d 1133, 1135 (Ct. Cl.1971) (per curiam); Louis Loss, Fundamentals of Securities Regulation 671, 792-93 (1983); Zvi Bodie, Alex Kane & Alan J. Marcus, Investments 89-91 (3d ed.1996). He is not paid by the stock market for this service, but is compensated by the income he makes from his purchase and sales and by commissions on limit orders (orders contingent on a stock’s price hitting a specified level) placed with him by brokers. United States v. Bleznak, 153 F.3d 16, 18 (2d Cir.1998); Fridrich v. Bradford, 542 F.2d 307, 309 n. 5 (6th Cir.1976); Loss, supra, at 671; Bodie, Kane & Marcus, supra, at 90. The Internal Revenue Service treats his gains and losses as ordinary income because the Internal Revenue Code classifies him as a dealer. See 26 U.S.C. §§ 1236(a), (d).

Treasury securities, at least the ones in which Bielfeldt transacted, are not sold on an organized exchange, and so there are no floor specialists — there is no floor. The market for Treasury securities is an over-the-counter market, like the NASDAQ. But the economic function that the specialists on the organized exchanges perform is independent of the form of the market, and dealers who specialize in Treasury securities (called “primary dealers,” and discussed in the next paragraph) are close analogues of the floor specialists, just as NASD market makers are. There is even a recent law that requires the primary *1038 dealers in Treasury securities, with some exceptions, to register with the SEC or the NASD. Government Securities Act of 1986, Pub.L. 99-571, §§ 102(e) and (f), 100 Stat. 3218,15 U.S.C. § 78o-5.

Bielfeldt claims that he performs this function too, though he is not a registered or primary dealer. The securities in question are used to finance the national debt. During the period in which he incurred losses, there was no talk of paying off the debt — on the contrary, the debt was growing. To finance growth and redemptions, the Treasury would periodically auction large quantities of bonds and notes, which would be underwritten by a relative handful of primary dealers. Bielfeldt would buy in huge quantities from these dealers and resell in smaller batches, often to the same dealers, a few weeks later. His theory, which worked well for a few years and then turned sour, was that the Treasury auctions were so large that each one would create a temporary glut of Treasury securities, driving price down. He would buy at the depressed price and hold the securities off the market until, the glut having disappeared (because he was hoarding the securities), price rose, and then he would sell. He argues that had it not been for this service that he performed in the marketing of Treasury securities, the price the Treasury got at its auctions would have been depressed, with the result that interest on the national debt would be even higher than it is.

What he is describing is simply the social benefit of speculation. Think back to the Biblical story of Joseph. During the seven fat years, years of glut, Joseph “hoarded” foodstuffs so that there would be an adequate supply in the seven lean years that he correctly predicted would follow. In a money economy, he would have financed the program by buying cheap, which would be easy to do in a period of glut, and selling dear, which would be easy to do in a period of scarcity and would help to ration supplies in that period. He would buy cheap yet pay higher prices than people who were buying for consumption, since he would anticipate a profit from the later sale during the period of scarcity. Similarly, Bielfeldt hoarded Treasury securities during the fat weeks immediately after an auction so that there would be an adequate supply in the lean weeks (the weeks between auctions) that followed. That activity may have been socially beneficial, as he argues, but it is no different from the social benefits of speculation generally. His argument if accepted would turn every speculator into a dealer for purposes of the Internal Revenue Code. United States v. Diamond, 788 F.2d 1025, 1030 (4th Cir.1986).

Unlike a floor specialist, Bielfeldt undertook no obligation to maintain an orderly market in Treasury securities. He did not maintain an inventory of securities; and because he skipped auctions that didn’t seem likely to produce the glut that was the basis of his speculative profits, there were months on end in which he could not have provided liquidity by selling from inventory because he had no Treasury securities. In some of the tax years in question he participated in as a few as 6 percent of the auctions, and never did he participate in more than 15 percent. As a result, he was out of the market for as much as 200 days a year. He was a speculator, period. As the Federal Reserve Bank of New York, which kept track of Bielfeldt’s trading in Treasury securities and sent updates to the IRS, put it, “his activities are in most cases outright speculation on interest rate movements.”

In saying that Bielfeldt was not a specialist, we don’t mean to imply that the Internal Revenue Service would be required to recognize as a dealer a trader who structured his operation to resemble that of a floor specialist but was not a floor specialist as defined in 26 U.S.C. § 1236(d)(2). That issue is not before us. Nor is the bearing of a 1997 statute, 26 U.S.C. § 475(f), which allows a securities trader to treat paper gains and losses as ordinary rather than capital income by *1039 marking to market the securities he owns at the end of the tax year, that is, by pretending they had been sold then. We note finally that Bielfeldt’s alternative argument, that Treasury securities are “notes receivable acquired in the ordinary course of trade or business” and therefore are not capital assets within the meaning of 26 U.S.C. § 1221(4), is frivolous. It implies that no bonds, government or private, are capital assets, since a bond, like a note receivable, is a promise to pay the holder of the instrument.

Affirmed.