Harberger Taxation

One suggestion is to extend the current practice of patent maintenance fees (and historical practice of copyright renewal), to a “Harberger tax” (a mechanism like a “shotgun clause”, recently re-popularized by Posner & Weyl2016): owners of copyrights must pay for each time period the copyright is enforced, based on its total value, like a fixed percentage of its value (eg. 5% annually). It is hard to know what the value is, so a good trick is self-assessment of the value, to enforce an honest valuation by the person who would best know the value, the owner: the owner provides an estimate of the value, and is required to sell to anyone at that value; if it is undervalued, then everyone else will try to buy it from them. And owners who ‘overvalue’ a copyright to ensure they keep it are merely demonstrating they value it higher than anyone else and are willing to pay for the privilege.

This is appealing (eg. it removes the strange subsidy of copyright being completely exempt from estate taxation, which is motivated in part by the difficulty of fair valuation) but has a few issues as a way to handle valuable copyright estates.

A Harberger tax on all works is potentially a major hassle compared to ‘born copyrighted’ for active, living authors. Just the need for copyright registration is a major problem. (Do I have to register every page on this site, or every Reddit comment I write? And I am punished for writing more by paying more, even if that writing is an asset?) It requires constant administration for all owners to keep updating the value and paying the new tax. (Or do copyrights expire the moment anyone files paperwork a bit late?) The constant risk of being bought out may discourage owners: even if we ignore the abuse potential as self-correcting (“please don’t throw me in that there briar patch by overpaying!”), the intrinsic instability and risk is a chilling effect. (Who would want to enter into a multi-billion-dollar/decades-long deal with, say, the Tolkien estate if it might change hands at any moment? Or want to write something that becomes valuable if they might lose it by not updating valuation fast enough—what happens if something goes viral and overnight is worth millions?) Proponents like Posner & Weyl2016 have answers to many objections to Harberger taxes, like software for adjusting valuations in realtime without human effort, which are plausible⁴, but they compromise the original simplicity of the proposal and it’s unknown how well they would work. There is also still no easy way for disgruntled fans to try to buyout the owners easily. (The benefits are too concentrated, and the harms too diffuse—any individual Dune fan disgusted by Brian Herbert may be willing to kick in $10 or $20 to take the copyright away from him, but they can’t do so for lack of coordination. Dominant assurance contracts struggle to fix this because it is not known in advance how much would be necessary.) And while the owners do get some steady encouragement to make good use of the copyright as the Harberger tax keeps nibbling away at them, but the opportunity cost remains invisible (similar to real estate owners who will squat on a property indefinitely as long as the low property tax can be paid out of rentals, and owning it is “free”), so we can expect many orphan or underused works still.

OK, so we exempt living authors entirely. Our primary concern is with posthumous mismanagement and long-lived works of commercial value, so we leave standard ‘life copyright’ alone; this is strictly a matter for estates or corporations etc which register & pay for specific IP. We know that that’s not a big deal for those entities, because registration is how copyright (and patents still) have worked for centuries (and they didn’t even have computers). Authors need not worry about the hassle—because they’ll be dead.

Now, as for the issues around Harberger taxes being blunt instruments, with appealing theoretical properties for Homo economicus but seeming rather bruising if applied to the real world (eg. “the billionaire buyout problem”). What if we reframed the Harberger tax as a kind of descending auction, where the tax starts at 0%, increases each year, and eventually just reaches 100% and the copyright expires? Then we have less orphan works than a fixed percentage, because of the ramp up. We continually squeeze works out into the public domain, so they only stay copyrighted if they are privately valuable.

But this would fail on a couple criteria, particularly by discouraging investment or exploitation: the better you do, the more you are punished, and you may have major problems with cash flow or unrealized gains costing you everything (eg. a major movie deal is signed, and the estate gains hundreds of millions of dollars in value that year, but the movies have not been made and so not a penny in royalties has yet arrived, even as the tax percentage also goes up). This also still doesn’t really fix issues like abusive but lucrative ownership.

Pricing Opportunity Cost

What we would like is some sort of mechanism which ties the Harberger tax & self-assessment to the opportunity cost, such as just investing the equivalent amount of capital in the stock market, which is more carrot than stick, so owners want to carefully track ownership and they want to sell copyrights if they are not enjoying very high returns or are going to passively manage them so they can enjoy a quiet life, and also make it easy for third parties to buyout or contribute to a buyout. How could we set that up?

One way we could do it is by thinking of everything in units of stock market indexes. If an estate throws off income of only 5% a year, but a stock market index is returning 7%, then we would like the estate owners to see that they are not doing well, but poorly, and underperforming by −2%, and we would like them to want to sell the estate, or failing that, just revert copyright to the public domain. You can see this just by comparing the statistics, of course, but that doesn’t provide an incentive. (Few owners will be so financially-minded & competent that the comparison would occur to them, much less motivate them—that’s the problem!) We have to tie the copyright to an actual index investment.

What if… the ownership… were actually of both a copyright and an index, somehow? And the owner had to choose which, and if they weren’t doing a good job or the copyright wasn’t valuable, they’d choose the index?

This wouldn’t force bad owners to sell (eg. Brian Herbert might refuse to sell at almost any price because ruining Dune gives his life meaning), but it will convince most bad owners to sell. It’s too bad that we cannot solve the holdout problem perfectly as long as we rely on voluntary sales—but there is no such thing as a perfect mechanism, and adding in pure Harberger-style forced-buyouts brings in too many problems.

I also appreciate the esthetic & moral dimension here: intellectual property is not real, and a government-enforced monopoly which takes from everyone in the world; as the US Constitution makes clear, there is no intrinsic or moral right to “intellectual property rights”, as they are a taking from everyone else, and justified solely by their pragmatic value to “promote the Progress of Science and useful Arts”. If rentiers who have bought or inherited a copyright then squat on & mismanage it, it is wonderful apt that their mismanagement is both self-punishing and repaying society—by forgoing consumption of the index in favor of long-term investment of the original capital, which increases economic growth, which indirectly “promotes the progress of science and useful arts”.

Standing Buyout Offers

Can there be some sort of government sinking fund for regularly buying out IP, similar to how governments sometimes buy out patents? Well, probably not, because that would incur all sorts of problems of deciding which to buy at how much, and where does the money come from, and how do you avoid political corruption—particularly if you are trying to use this in a major IP-heavy economy like the USA?

Still, there’s something to this notion… What if we instead had buyout funds per estate? The government might, say, put in an endowment of $Z into a stock market index investment account, legally earmarked for buying out a particular estate, and which compounds indefinitely. (Should the copyright expire without a buyout, not even at the last second, then the government just keeps it as general revenue.) Stock market indexes are so large that they cannot be manipulated by any individual owner, while following objective rules (the stocks chosen by an exchange) and being extremely cheap & able to handle millions of accounts corresponding to billions or trillions of dollars. And since the index exists to eventually buy out copyrights, but not at any given time, market volatility doesn’t matter.⁵ (Estates may think they can ‘time the top’ and ‘exploit’ the indexes by timing their redemptions, but since markets are efficient, they are no more able to do so than any other trader, and probably less.)

The estate may refuse initially, but as time passes and the investment compounds, if the estate is doing a poor job, the investment will grow to become arbitrarily larger than the estate, and the estate will be desperate to take the buyout because of the opportunity cost. The opportunity cost isn’t invisible, but as clear as the S&P 500 ticker scrolling along the bottom of CNN. (Indeed, they will want to take the buyout once the ratio starts changing, because they could just leave the index where it is and enjoy those greater returns.) And then anyone who wants to contribute to the buyout has an easy way to do so: just deposit some money into the investment account, where it becomes part of the earmarked buyout, and assists the exponential growth. (No need for any paperwork or permission—just send the money!⁶) This is efficient because the endowment is invested well by default, and indexes are super-cheap to buy or administrate (and get cheaper the more is put into them); indexes also generally do not risk serious political corruption or economic distortion, and can scale to trillions of dollars.

This is also efficient because the value of each account can be easily publicly released (as well as calculated from the initial endowment + known index returns), so everyone can see exactly how much the buyout for any given estate is—which provides a lower bound on the current value of the estate, inasmuch as otherwise the owner is leaving money on the table by still refusing the buyout. This greatly reduces the risk of attempting to negotiate a deal or executing a deal (the index puts a floor on the estate value, which greatly increases its total value); one can even make standing offers like “I’ll pay you the index value plus 20%; call me whenever you decide you want to stop being poor”, reflecting the control premium or returns to better management.

This blind offer capability also helps negotiate with orphan owners, preserving the ability to take a finder’s-fee without tipping one’s hand for free: you can track down owners of valuable estates, who don’t know they own it, and make them a provably-fair offer like “50% of the index value”, which is objectively verifiable after the deal as a fair valuation, and so the owner can agree to it without even knowing what it is. Otherwise, no deal would take place—of course if someone came to you offering you some money for a copyright you didn’t know you owned, and they won’t tell you what the copyright is because then you would just go and sell it to someone else, you would refuse! Because how do you know they aren’t ripping you off by offering chump change? You would figure it out after the deal was executed, of course, when you found out exactly what copyright you owned (left to you by a distant great-uncle’s niece you had no idea had become world-famous for her comic books set in 1800s Russia about dancing circus bears), but that would be too late. This makes markets in information difficult and tending towards lemon markets and killing the possibility of transactions.

Well, that sounds nice, doesn’t it? What owner would complain about voluntarily selling their copyrights for an ever-increasing payday? Or having many viable offers to sell at higher prices, should they wish to? Or being able to tell disgruntled fans, “if you don’t like it, help buy me out and show you can do it better”? (It also seems mostly compatible with existing IP law like the Berne Convention: nothing in it exempts copyright from taxation, and any owner can still run out the full copyright term if they want to enough.)

Self-Funding Harberger Tax Buyouts

Then the rest of the scheme follows as before:

1. all existing copyright is grandfathered in and operates under the old system; the new system begins at a certain date

2. Estates with minimal commercial value choose to let works fall into the public domain immediately by default, in an easily-verifiable fashion (no registration in the copyright registry), enabling all of the valuable uses of the public domain and positive externalities

3. Estates with commercial value instead choose to declare their best estimate of the value, with honesty enforced by a one-time-only mandatory sale during the probate period

   • and X% of that declared value is ‘taxed’ and invested in a S&P 500 stock market index or other broad stock market index, and can only be spent to buyout into the public domain the associated estate; this avoids wasting capital and increases economic growth

   • Anyone can at any time donate any amount to the buyout fund

   • The current buyout balance amount is always publicly available

   • Ownership can be fractional: if you want to purchase a copyright from a owner, you are also purchasing an associated percentage of the index and splitting it (which may be 0%); as the index is so straightforwardly valued by ordinary accounting methods, and has no especial value on its own and is independent of the copyright’s success or failure, this presents no difficulties.

     Derivative works are just standard license contracts + new copyrights. (If the original license becomes invalid because the copyright owner cashes in the index and the copyright falls into the public domain, then that also means there is no need for the license.)

4. Estates can at any time accept the buyout, or sell to a better offer (enabled by the pricing information and a much more liquid IP market growing up, backstopped by the indexes). These sales are always voluntary.

   • speculators will be enabled to look for undervalued works (to better manage) and featherbedding management squatting on estates rather than admit their incompetence. (From one perspective, an estate licensing a lot of stuff can resemble a closed-end mutual fund…)

   • “bounty hunters” are incentivized to hunt for quasi-orphan works, which are either underused or should be cashed in (removing their orphan status and reducing the deadweight loss), as a sub-market enabled by objective third-party valuation

   • dominant assurance contracts also become less of a gamble: even if they do not immediately succeed in the buyout, they will still increase the buyout fund closer to the critical level

5. Copyrights terminate after a large fixed term if they still exist, and the index reverts to the government (or charity etc, like most abandoned or untraceable property)

6. work-for-hire copyright / sales of copyright, with, by definition, a fair market price set by that transaction, skip the self-assesment mechanism but otherwise operates similarly:

   either public domain immediately, or copyright registration requires X% to self-fund an index.

Other Uses

While I thought about this in the context of long-lived copyrights & franchises, perhaps it could apply to other things. The key traits seem to be: long-term hard-to-value illiquid assets with invisible opportunity costs & externalities/spillovers but where management can make a dramatic difference to the final realized value.

Other forms of intellectual property like patents are obvious examples, where there is a “public domain” for them to revert to; but this also describes, to some extent, real estate, mineral rights & water rights, radio spectrum licenses, and Internet domain names.

I’m less sure if it could be a good fit for various kinds of renewable or inelastic quotas like fishing quotas, carbon emission rights, occupational licensing like taxi medallions, fine art, or software platforms. These generally are more conserved or transferable or lack externalities/spillovers: fish don’t vanish if a fisherman doesn’t use their quota, nor is there much difference in how one catches a fixed number of fish, and you need a certain number of taxis but you can just tax them normally due to their fungibility (and set the tax rate by auctioning off taxi medallions, which you need to do to handle turnover or population growth anyway). There’s usually not much need for a Harberger tax there, much less one with any exotic ‘self-funding’ mechanism.

External Links

• Discussion: HN

Similar Links