SWF+UBD

Matt Bruenig has published an excellent proposal that the United States charter a large sovereign (“social”) wealth fund and use its profits to fund a universal basic dividend. It has provoked a flurry of thoughtful responses, see Mike Konczal, Matt Yglesias, Peter Gowan, Owen Davis, Peter Barnes, and Kevin Drum. (Bruenig has published a response to Konczal. I’m adding links to further discussion in an update section below.)

I. Could a SWF fund a UBI?

As regular readers (of this irregular writer) know, I enthusiastically favor a UBI. A “universal basic dividend” is not quite the same thing. The idea in a nutshell is that the Federal government would charter an investment fund, of which each adult US citizen would be granted an share. The shares could not be sold, redeemed, or bequeathed, but they would pay a dividend from investment proceeds. The government might seed the fund by selling underutilized land or assets, or by a simple grant financed by the Treasury. Thereafter, new contributions would be made from a variety of dedicated tax streams and from the compounding of undistributed earnings. Initially the dividend would be quite small, but over time, it would grow.

Would it eventually become a “full” basic income, covering the $1K-ish per month UBI advocates often kick around? Maybe, but not for decades. Bruenig proposes paying out roughly 4% of fund value annually, assuming total returns would generally be higher than that. To generate a $12K per year UBI for US adults in 2018 dollars, the fund would need to grow to about $75T. (Inflation and population growth will increase that gross number, but they will also increase the scale of contributions, so I think that for an intuition that’s a good start.) I mistrust aggregate balance sheets — I think they are largely meaningless — and I especially mistrust the Fed’s Flow of Funds Table B.1. Derivation of US Net Worth, which I consider incoherent. Nevertheless, on the presumption that a canny astrologer is superior to no lodestar at all, we can take for scale the Fed’s B.1 estimate of US Net Wealth — $93T — and its (less incoherent) B.101 estimate of total household and nonprofit sector assets — $116T. (Household assets indirectly include business sector assets, since households own businesses.)

Basically, on these numbers, the fund would have to grow to hold something like 64% of all assets, or 80% of US “net worth”, to finance a “full” UBI at a 4% per annum payout rate. That sounds… drastic. But it should be taken as a flawed baseline to start thinking from, not as the final word on anything. In particular, the value of financial capital is largely institutionally determined. In a world where the state largely corners the market on it, do capital prices rise as private sector actors compete for the crumbs the solidarity funds leave behind? If so, the share of ownership required to fund a $12K UBI would fall, somewhat. But if, as in theory, financial asset production is price elastic, that same $75T could become a much smaller share of total assets (though not necessarily of net worth) as the financial sector produces new assets to satisfy demand without expanding valuations. Would public ownership via the investment funds cause businesses to behave in ways both less predatory and less profitable? Then the total market value of assets might deflate, leaving the share necessary to fund a UBI even larger. Alternatively, public ownership could cause firms to reign in exorbitant CEO salaries, payouts to well-connected professionals (lawyers, financiers, consultants, etc.), and shearings by insiders and activist investors, rendering them more profitable to long-term shareholders, and rendering firm shares more valuable. This would reduce the share of assets that would be required for a UBI-sized dividend. Note that it is entirely possible for the fund to grow larger than the total value of US assets — it can (and should) diversify into international assets as well, which should become more plausible over time as the United States’ share of global economic activity is likely to decline. However, if the program is successful in the US, it’s likely that “social welfare funds” will explode from their current Nordic ghetto and become widely adopted by wealthy countries. So, “in equilibrium”, we should expect most financial assets to be owned by these funds, if we mean for them to finance anything as ambitious as a UBI.

One important lesson of our little thought experiment is that a straight-up $1K/month UBI — the kind I am glad to propose and advocate, that would be financed conventionally via tax, debt, money, or cuts to other spending — would never be mostly a transfer from “capital” to “working people”. We live in a society with tremendous inequality in what gets scored as “labor income”, but what is really returns due to the scarcity of very particular institution-specific skill sets, or monetizations of social and political connections, or all kinds of other things many of us would code as rent-extraction. My first framing of UBI is as a “fixed/floating swap on highly variable income”, most of which we call wage income, not capital income. An income-tax financed UBI is a form of insurance in which people give up some of the upside on their potential earnings for a fixed income that can serve as a base. Ex ante it is insurance, ex post it becomes a redistribution over all income, not just from “capital”, which makes it much more plausible to finance. “Labor” (much too broadly defined) still constitutes the majority of total income. Financing something as big as a UBI only from the income we code as “capital” means taking most of that. Also, it’s worth noting that even if our SWF grew large enough to generate dividends at the scale we might desire for a UBI, a universal basic dividend would be less effective as insurance than a fixed, Treasury financed UBI, as dividend payments would be be “procyclical”, decreasing when the economy does poorly and recipients are most likely to experience shortfalls of other income. Bruenig’s proposal mitigates this somewhat by proposing the dividend rate be based on a 5-year average asset-base, rather than fluctuating with a single year’s performance.

There is no need to make a “full UBI” an arbitrary lodestone or hurdle for an SWF dividend. The Alaska Permanent Fund pays out roughly 10% of what I am calling a $12K/year “full” UBI, and it is an enormously successful program. A $100/month UBI might not sound like much to most of my readership, but it would completely end $2/day cash poverty in America, and would be a huge deal to those who need help most. A UBI means equal payments in dollar terms, but is highly progressive in terms of welfare as commonly modeled, as the poor get a much bigger benefit from a small income than the wealthy do. If we are collectively uncomfortable with the degree of public ownership implied by a $75T social wealth fund, we could rather quickly (in about a decade) ramp up a fund that would generate an Alaska-sized payout. Doing so would not foreclose the prospect of financing “the rest” of a UBI, or a job guarantee, or anything else via other tools. The only constraint would be “fiscal space”. But a strength of the social wealth fund structure, perhaps its main strength, is that it may create new political space for redistributive taxation.

II. On the political effects of a SWF+UBD

“Taxing the rich” polls well. But “sending your money to Washington” probably doesn’t. A clear advantage of a social wealth fund attached to a citizens’ dividend is that it makes a great destination for potentially popular redistributive taxation. It’s the usual bourgeois morality tale about saving — start with a little, keep contributing, watch it grow, enjoy pride of ownership — as a national project. Dedicating tax streams to the SWF neutralizes objections that “Washington politicians will just waste it”, assuming the fee ratio can be kept low. Initially the dividend will be small — for the first few years it should probably just be zero — but each citizen would have her notional net-asset-value that she can watch grow. (Bruenig proposes that there literally be an app for this.) We will all have cause to cheer a good year on the stock market.

Which brings us really to the bloody heart of the controversy that surrounds this proposal. Maybe it is true that a dividend-devoted SWF could render redistributive taxation more palatable, but mightn’t the same enthusiasm trick the working class into acquiescing to predatory corporate behavior in the name of increased profitability? This is Mike Konczal’s most serious objection. If we imagine an Alaska-scale SWF, you can argue that this proposal would be a terrible own-goal for those of us concerned with remedying social stratification. The rich are always trying to persuade the rest that what’s good for the stock market is what’s good for America, what’s good for us all. That’s mostly a lie. Decimating labor unions was good for stock markets, good for profitability, but bad for any hope of an equal society. Fracking every goop of hydrocarbons out from the Earth might lead to outsized S&P 500 performance for a few years, but would end up being a bad deal for most of us. Concentration and monopoly power are good for profitability, but are terrible for almost everyone in their roles as consumer and worker, and are destructive of economic dynamism and growth. If everyone suddenly sees themselves as a shareholder — while the true benefits of shareholding continue to go disproportionately to the affluent — then the plan just becomes a warmed-over version of George W. Bush’s “ownership society”.

Bruenig’s rejoinder — also serious — is that when “capital” becomes the broad public, it will recognize its own interest and cause firms to behave differently. An Alaska-sized fund, while it might only hold 6% or 7% of total US assets, would likely hold disproportionate positions in the larger, more liquid, firms that dominate the economy. “The public” would then be a 10% or 20% shareholder in blue-chip megafirms. And given the dispersion and passivity of shareholders in large public companies, a 10% to 20% interest can exercise considerable influence over the operation of the firm. So we have to consider a balance of two opposing effects: From the outside, the broad public might identify more with business interests, which at this scale would not meaningfully be its own, and so become less sympathetic to labor, environment, anti-trust, social-justice, and other forms of activism that might otherwise hold firms to account. But from the inside, the broad public would have more direct influence over corporate behavior. Which of these effects would dominate is, I think, anybody’s guess.

If the SWF grows towards full-UBI scale, it ceases to be meaningful to talk about an interest of “capital” distinct from the broad public, or at least distinct from the broad public as constituted by the management of the SWF, via legislative mandate. In theory, the conflict between capital and workers will have finally been resolved, as the workers and the capital owners will be the same people. But, in practice, whether big or small, how would the SWF vote its shares?

A lot of us, I’d venture to say most of us, are pretty cynical about the capacity of our “democratically elected government” to represent even a loose caricature of a general public interest. We presume our representatives and regulators are usually bought, or captured by lobbyists with gentle words that sound like sense but are really industry dollars. We presume they are seduced by revolving doors and the demands of a donor class. Do we really want these slimeballs managing — potentially mismanaging — our industries?

Bruenig, in his proposal, has one smart response to this:

It is fair to worry that the government might make bad shareholder votes from time to time, but not reasonable to think that very affluent people will on average make better shareholder votes than a democratically-elected government.

Put a little more cynically, the government might do a shitty job of representing a broad public interest, but current shareholders aren’t even trying, or pretending to try. They are voting outright the narrow interests of the those whose corruption we fear. Whatever echoes of a general interest find their way through all that corruption would represent a step up from the status quo of explicitly plutocratic corporate governance.

There’s another take on this that I think comes through in Kevin Drum’s cri de coeur. Instead of contrasting a (good) public interest against a (bad) plutocratic interest, one might instead worry over the distinction between (good) commercial interests and (bad) political interests. In one mythology of capitalism, one that I myself have held dear, the secret of our collective economic intelligence is that businesses are governed by commercial incentives and constraints and disciplined by the prospect of failure. Political bureaucracies, on the other hand, face incentives and constraints that are largely endogenous to how they happen to be constituted, from the demands of always somewhat ridiculous electoral contests to accidents of who happens to know whom and how favors are traded. The coupling between political choices and the quality of actual outcomes is squishier than a spent condom, and always refracted ex post through a capricious contest over who must be blamed. Ultimately, in this view, the behavior of such bureaucracies is mostly arbitrary with respect to the actual outcomes and interests to which they are putatively devoted, and can only really be understood in the context of parochial circumstances and interests within, and the outside forces that strive to shape those.

In this account, Bruenig’s story about democratic representation being at least a step up from outright plutocratic control makes a category error. “Democratic representation” wouldn’t be that at all. It would just inject into business management a kind of political id, tangling the resources of productive enterprise in politically faddish boondoggles or the corruption not of monied interests but backscratching self-important mandarins. In economist dork-speak, the trade-off between plutocratic and broader public interests might constitute a kind of Pareto frontier, and we might wish to shift to a place on that frontier that traded away plutocratic interest for more of the broad interest. But throwing political bureaucracies into the mix wouldn’t execute that trade. Instead it would drag us into the chill interior, where neither party’s interest is efficiently served and potential welfare is left to rot while the betentacled idiot blathers and drools.

I have some sympathy for this account, but I no longer consider it as disposative as I would once have. First, it is clear to me that very large firms bear a great deal of resemblance to the political bureaucracies that their captains so malign. Large firms, just like states, are run as bureaucracies, and the people within them are governed in a day-to-day way by the same sort of careerist and interpersonal arcana that drive public bureaucrats. Large firms very frequently do not face sharp and immediate commercial constraints. How many years before it succumbed to formal bankruptcy did General Motors truck on with negative book equity? And then, of course it was disciplined by the terrible indignity of a bail out. Obviously, large banks have been proven to face what economists euphemistically refer to as “soft budget constraints”. How much money does Google — excuse me, Alphabet — waste on the bizarre pet projects of its increasingly extraterrestrial founders? Where is that vicious market discipline? Of course, stock markets could turn on Google — excuse me, Alphabet — at any moment, and the firm’s managers (who do not actually face the bracing discipline of potential bankruptcy on any bureaucrat’s relevant timeframe) could suffer some embarrassment. This strikes me as analogous to the risk faced by politicians that some gaffe or issue upon which they are poorly positioned will subject them to the erratic ire of the media-activist blob and jeopardize their electoral prospects. The governance of large firms and states is more alike than different. In both cases, they effectively exercise power and achieve important goals not from some extraordinary capability of decision-making, but by virtue of extraordinary resources that render mistakes survivable and enable them to succeed despite the pathologies that come with management at scale.

If large firms are like states in their (non)quality of decision-making, they are also like states in another way. Their choices provoke consequences that seriously affect the welfare and interests of large groups of people — workers, customers, vendors, taxpayers, breathers. It may be tenable to claim that small firms are, to a first approximation, purely private affairs. But the choices of Facebook and Google, Exxon and Ford, can affect the welfare of far-flung and indirectly connected stakeholders more powerfully than most decisions of sovereign states. As a practical matter it may be difficult to arrange, but as a substantive matter, the cause of welfare demands representation by broader stakeholders in large firm governance for precisely the same reason it demands representation within states. Without such representation, important interests will be overlooked. In some imaginary world of myriad firms tightly constrained by competitive markets whose individual pursuit of profit would as if by an invisible hand find equilibrium at general welfare, this sort of representation would be unnecessary and superfluous. We do not live in that world.

Libertarians do object, but this logic is perfectly conventional in practice. Many kinds of firms face increasing regulatory scrutiny as they grow large. One might ask (as I think Mike Konczal implicitly does) why formal public representation should be necessary, since after all the state already has this means of exercising control — regulation — which requires no stock ownership or board members. In an idealized world, it would not be necessary. Regulatory powers leave the state with extraordinary ability to exercise control over firms.

But in reality, regulatory power is insufficient. If you are concerned about the informational problems associated with the state mucking around with commercial decision making, you should be even more concerned with regulatory control, as regulators make broad choices that affect whole industries under circumstances wholly divorced from context or specificity of circumstance that might in practice render those choices ineffective or counterproductive. Because of this lack of context, regulation should and generally does restrict itself to imposing guard rails or broad mandates. In practice — from the perspective of the broad public, not just whining firms — the exercise of regulatory authority is costly, and states are wise to use that authority warily. But that leaves a great deal of space where a public interest would like to assert itself, but for which regulation is too blunt an instrument.

Also, for all the imperfections of the US Congress, the regulatory state is much more plutocratic than the legislative state. Broad public interests are most likely to have a shot when an embarrassing public spotlight is thrown on legislators, when well-publicized votes are token by the most visible rule-making bodies of the land. Narrow interests capture the regulatory state because it is too vast and boring for the improvised supervisory capacity of diffuse interests to monitor or contest. Regulatory vigilance does not survive what Minsky called “periods of tranquility”. Concentrated interests are relentless, diffuse interests depend fatally upon the caprices of the news cycle. Captains of industry seem to hate regulation, because they are bound to hate whatever regulations find favor in the public’s eye. They are not so deregulatory in the shadows. Quite the contrary.

This is why we suddenly find ourselves talking about “codetermination“. Where regulation would impose course-grained constraints on actors that we hope will yield better overall outcomes, representation of new interests on corporate boards changes the nature of the actors themselves, in hopes that their interests will become more aligned with a broader group of stakeholders. Bruenig’s share-voting SWF would represent a slow transition to codetermination in another form. Worker democracy is imperfect, but looking to Scandinavian countries or Germany we may become persuaded that election of workers to corporate boards is valuable, even though worker interests might in theory have been asserted through labor regulation instead. The same might be true in America, not just for workers but for the broad public.

I find myself surprisingly comfortable with these ideas (once I would not have been), so long as they are restricted to larger firms. For smaller firms, there really is a “because freedom” objection. Large, public firms are run by officers appointed by a committee elected by a population. Changing the composition of that population alters the purposes to which the firm and its officers bend themselves, which were never rightly their own. It restricts no one’s freedom. The actions of closely-held firms might seem extensions of the will of a small number of people, and those people certainly might consider being forced to cede that perfect control an abridgment of their freedom. But claims to liberty can only be accommodated to the degree to which they don’t infringe upon the rights of others. The larger a firm becomes, the more market power it achieves (among consumers, vendors, or workers), the more its choices affect the welfare of others. Libertarians with their fables of procedural legitimacy hate to acknowledge this, but there is no bright line between what is voluntary and what is coercive, and all of us, as individuals or businesses have some capacity to coerce. In the name of liberty, we must overlook that when the scale of potential coercion is small relative to the actor’s liberty interest. But it is also in the name of liberty that we must not fail to note when the scale of an actor’s ability to coerce is large so that their actions implicate the liberty interests of many others. Those who value their own liberty above all else must tolerate modest scale and power. Those who wish to captain large and powerful firms must tolerate some abridgment of their liberty to do with those firms as they will. Whether in the public or the private sector, there are and should be trade-offs between power and freedom. Power demands constraint.


We began this section by pointing out how an SWF could create political space to enact more redistributive taxation. Let’s end by pointing out something a bit less obvious and a lot more powerful. A sovereign — er, social — wealth fund is a taxation machine. It is an automatic taxation monster. It takes the miracle of compound growth that capitalists are always on about and turns it into a miracle of compound taxation, effectively taxing wealthier cohorts (those who would otherwise own the SWF assets) an ever increasing share of income year after year without requiring any new legislation, and with minimal distortion of investment behavior.

To see how this works, let’s imagine that we want to simulate the flows of an SWF+UBD. We’ll imagine a very simple scenario. Let’s define a “notional” SWF. The SWF is going to be financed by a tax enacted just once, which will yield $1T in Year 0. The tax take will grow with nominal GDP, which we will model as growing at 5% annually. Beginning at the end of Year 1, the SWF will make payouts. For simplicity, we will base payouts and returns on the end-of-prior-year balance. That is, we are conservatively assuming that the taxes we collect within a year are unavailable until the year following. We will assume a constant rate of investment return of 8% per year. Echoing Bruenig’s proposal, we will have the SWF payout 4% of the prior year balance each year.

However, instead of actually forming the SWF, let’s say that the government were to decide that there’s no need to intervene in the miraculous private sector with actual state ownership, that the assets can remain, um, efficiently managed in private hands but the government will simply use the tax system to reproduce the flows an SWF would generate. As it would if it actually formed the SWF, in Year 0 it would enact a tax, which would raise $1T. At the end of Year 1, it would have raised an additional $1.05T from the same tax. The notional SWF would have enjoyed the same $1.05T a new contribution. However, the notional SWF, if it had actually been constituted, would have also earned $0.08T as investment returns. In order to simulate the SWF flows, the state would have had to adopt a new capital tax of $80B. In Year 2, we have the same effect again. The originally enacted tax now raises $1.1025T, and the new Year 1 tax brings in $84B (assuming that both grow in line with GDP), for a total intake of $1.1865T. However, investment returns on the prior year SWF balance of $2.09T would have yielded $167.2B, which when added to the same take of $1.1025T from the initial tax, yields an inflow of $1.2697T. So, to bring the total inflows in line with what an SWF would have done automatically, the government would have to impose a new tax of $83.2B.

And so on. Each year, to reproduce the same net flows from private capital holders as would “naturally” have occurred had there been a SWF, the state would have to enact a brand new tax, in addition to still collecting the taxes enacted in prior years. Under our assumptions, each year’s new tax is slightly smaller as a share of GDP than the prior year’s, but in reality, that would depend upon investment returns, gdp, and dividend payouts. Whenever investment returns net of dividend payouts exceed GDP growth, the effective new tax that would need to be imposed on capital holders becomes larger as a share of GDP than the prior year’s tax. (Here’s the little Mathematica simulation the numbers in this section are drawn from: [pdf][nb])

We can all, as Mike Konczal put it on Twitter, “spitball” about politics. But I think it fair to say that it would be difficult to sustain the political will to cumulatively impose new taxes on capital holders, every year, year after year after year over a period that might span decades. But the “gimmick” of actually using the proceeds from a single tax, enacted once and continued indefinitely, to purchase capital assets, generates the same effect as this compounding tax schedule in a way that seems natural and inevitable and legitimate under the norms of present-day capitalism. If we accept that other capital holders get to enjoy the miracle of compound returns, why shouldn’t a fund owned in equal shares by all citizens get to enjoy the same? Actually constituting a SWF delivers a regime of effective taxation that, I think it is fair to say, ordinary politics simply could not.

Further, the effective taxes embedded in an actually constituted SWF are efficient, in the usual economic sense of not distorting anyone’s behavior. If we actually tried to impose new capital taxes every year to simulate the inflows of a SWF, capital holders would be scrambling to find ways of squirreling away funds that evade the ever increasing taxation. With an actual SWF, only the initially enacted taxes that finance annual contributions are potentially distortionary. The investment decisions that would be distorted by the annual additional taxes are made directly by the SWF, which does not face any tax. The reduction of potential distortion achieved by actually constituting an SWF is considerable, as the cumulative magnitude of these additional taxes eventually overtake the annual contributions tax (after about 15 years under our assumptions) and continues to grow indefinitely. (While the fund does not face tax distortions, it invests as an agent of its citizen shareholders, so we do have to worry about managing agency costs. But that is a problem for most private investment as well.)

III. On the macroeconomic implications of a SWF+UBD

In Part I of this catastrophe, I considered whether a SWF-financed universal basic dividend could finance a UBI. Implicit in that discussion was that having a SWF accumulate 64% of the nation’s assets would be a heavy lift — “drastic”, I called it. But we’ve failed to really consider Bruenig’s proposal until we acknowledge that, for his purposes, this kind of scale would be a virtue, not a problem. Bruenig is, as we all should be, concerned about the incredible inequality of wealth in the contemporary United States (and, indeed, in most countries, including the Nordics which are much more wealth-unequal than they are income-unequal). Bruenig’s reasoning is pretty straightforward — if the SWF comes to hold all the wealth, and if every citizen has an equal share of the SWF, then wealth inequality is pretty much over. Of course he does not claim that the SWF would displace all private ownership of everything. But, as a matter of arithmetic, the greater the fraction of total wealth that is held by the SWF, the less inequality there will be for any given distribution of the remaining wealth.

As we indicated previously, however, there is no reason to think that the value of total assets would be invariant to the existence of a giant sovereign wealth fund. “Value” is remarkably protean. It is not some stable stand-in for anything we might understand as real wealth. So-called crypto assets have a market value of $195B as I write, yet a lot of economists would argue that their real economic value is decidedly negative, since they stand in for no obvious real economic resource and cause the waste of tremendous electricity. (FD: I work in crypto and don’t think this.) Crypto is not unique in having no fixed measure for its price. As Steve Roth has shown, valuation increases that can’t be mapped onto any conventional accounting of real resources dominate other forms of ex nihilo financial asset creation (like new government paper or bank lending) and account for between 15% and 25% of comprehensive income year after year.

Ceteris paribus, as the economists like to say, if a public investment fund came into asset markets with new money, much of that purchasing power would go towards bidding up the price of productive real assets rather than to efficiently capturing them in the name of equal citizens. That’s a problem for Bruenig’s story in two ways: First, the more “richly” assets are purchased, the less likely they are to be retain their value and perform well going forward, impairing future dividends. Secondly, if the new purchases are matched by asset appreciation, they will be less efficient at taking a share of national wealth and so reducing wealth inequality. Suppose total assets are worth $100 today, and the SWF comes in with $20 hoping to buy a 20% share. If those purchases drive the total value up to $200, however, the SWF obtains only a 10% share of (ex-SWF) national assets. The work of reducing wealth inequality has been significantly thwarted.

The lesson from this thought experiment is simple, and one Bruenig is likely to find congenial enough. To the maximum degree possible, purchases by the SWF should be matched by policies that diminish other flows into capital. To put that more straightforwardly, the SWF should be financed via taxes on flows that, if not taxed, would have been devoted to the purchase of financial or capital assets. This has a triple benefit: assets are acquired, the prices at which they are acquired is advantageous and likely to yield attractive returns, and wealth-inequality-generating purchases by private parties are displaced. Bruenig proposes a variety of revenue sources for a SWF, and many of them are rather ostentatiously levies on “capital” in a political sense, taxing activities of the financial sector like trading, M&A, and funds management. Maybe that’s a good idea. There are arguments for things like a “Tobin Tax” quite apart from the revenue it might raise. However, economically speaking, there is no need to restrict oneself to this kind of “look ma, it’s capital!” finance. Nearly all of very high incomes goes into purchasing financial assets and capital goods rather than financing consumption. Straight-up progressive income taxation — say a new high-income tax bracket or a surcharge on incomes above $1M — would be a fine source of revenue for a SWF. As we saw in Part I, distinctions between income derived from “capital” and “labor” are increasingly arbitrary, and restricting revenue sources to things clearly labeled “capital” is limiting from the start and will provoke entrepreneurial relabelings, like the “carried interest” loophole that asset managers famously use to get their income relabeled as tax-advantaged long-term capital gains.

However, some of the sources of finance that Bruenig proposes, such as using government debt or monetary seignorage to buy assets, might be counterproductive to the goal of reducing wealth inequality. These would constitute new rather than displaced money vying for financial assets, which would almost certainly bid up prices of existing assets or provoke the issuance of new assets. Particularly if existing asset prices are bid up, it’s not obvious that this wouldn’t exacerbate rather than reduce wealth inequality, given the extreme concentration of asset ownership and how SWF purchases might reshape the distribution of ownership. Trying to time asset purchases countercyclically might seem to mitigate this concern, but even there SWF purchases would be helping sustain wealth inequality that market movements themselves might otherwise undo. At any time in the cycle, purchases financed by taxes on funds that would otherwise go into assets would yield better pricing for the fund than money or debt finance.

Beyond reducing wealth inequality, how else would private-capital-displacing social wealth funds affect the economy? There are a lot of big questions here. After all, capital is more than moneybags to its owners. It is also the seed-corn for new economic activity. Historically we’ve relied on a mix of grass-roots bootstrapped risk-pooling, bank finance, and wealthy private capitalists to finance new businesses. This proposal would, by design, make the wealthy private investors ever scarcer. Bank finance, in the United States at least, has been eviscerated as a source of capital for entrepreneurial ventures whose assets do not serve readily as loan collateral. A combination of industry consolidation and regulatory constraint has rendered “soft-information”, “relationship” lending increasingly risky to, and rare among, bank decision-makers. Ideally banks are cooperative enterprises through which communities undertake the risks of their own development, but American banking institutions are simply no longer suited to that model. If we undertook to replace private capital with public wealth funds, the question of entrepreneurial finance is one we would have to design new institutions (or remedy old ones) to address. But we need to do that anyway: The existing system, under which capital allocation decisions are made disproportionately by a smallish number of segregated, socially homogeneous rich people is bad, both ethically and economically. Get on the right TED stage and billions will flow to your fashionable tech venture. But who will finance some business in Peoria that could absolutely be profitable in that very local context, but that would not much make sense to the globetrotters at Aspen Ideas? One answer is that the proceeds of a universal dividend, once expropriated from the faddish elite and spread widely among the public, might find its way to “crowdfund” local business everywhere. But we still have very few institutions that seem able to facilitate this sort of development. (This is one of the reasons why I work in crypto these days.) In my view, grassroots small-business finance is already an urgent problem. It would become more urgent if we eliminated the quixotic plutocrats to which entrepreneurs currently appeal, but replaced them with a bureaucracy of index fund managers. “Social wealth funds” should concern themselves from the start with questions of capital development, not view their mission as mere allocation among a menu of pre-existing public assets.

Conclusion-ish

Constituting a SWF, like organizing a job guarantee, is one of those ideas that could be wonderful or terrible depending on the implementation. It is easy to imagine the management of such a fund becoming captured by the financial industry. I am sure that clever financiers could devise compelling arrangements and accounting standards under which the formal fee ratio paid by the SWF would be lower than a Vanguard index fund while the public’s returns are quietly bled away by predatory intermediaries. Bruenig’s “American Solidarity Fund” could become a cesspit of corruption. To prevent that, it would have to be managed in an exceedingly transparent way, and should probably be organized with a great deal of internal competition. Balancing the vigorous policing of corruption against the need to accept growth-promoting risk — which means tolerating failures that ex post one might argue were ill-conceived and perhaps corrupt choices to begin with — strikes me as challenging. (Remember Solyndra?)

Further, as Mike Konczal emphasizes, a SWF is a long-term bet for a country that has acute short-term problems. If constituting a SWF were to occlude priorities like Medicare For All or ensuring full employment (via a job guarantee or other mechanisms), that would be bad and probably not worth the trade.

But there is no reason to think that it must displace other priorities. It is likely to create new space for redistributive taxation rather than vying for a fixed-size pie. In principle, a SWF+UBD has some extraordinarily desirable qualities. It turns the very mechanism by which capitalist dynamics concentrate wealth into a few hands into a means of disbursing wealth widely. It begins small and grows over time, which gives the citizenry and political system time to develop the institutions necessary to manage it (and leaves it disciplined, in its early years, by the risk that a hostile administration will liquidate it). It does not require wholesale reorganization of the everyday workings or ideological underpinnings of present society. Small-scale free enterprise and the contest for greater affluence remain intact, but the range of outcomes becomes compressed. There can be fewer billionaires if there are not so many billions of private wealth in private hands; there can be fewer desperately poor if we all derive some income from the collective trust fund. I think that this sort of compression, but not elimination, of variation in economic outcomes is desirable. It maintains incentives to produce, as economists conventionally understand them, while not allowing those incentives to become so extreme, so desperate, that they become incentives to cheat, to steal, to prey on one another.

Plus, as Matt Bruenig is at pains to point out again and again and again, social wealth funds are not some untried thing. They have worked pretty well for Norway and Alaska. Although you might object these are special cases because of oil revenues, the inequality-reducing capital accumulation machinery of SWFs will function regardless of source of funds. Overall I think this is an idea worth supporting.


Update: Here are two critiques of Bruenig’s proposal from a socialist perspective: “‘One Weird Trick’ to Building Socialism” by Frank Little and “A Quick and Easy Way to Forget about Socialism. Thanks to @BlueIvyRedWorld on Twitter for pointers to both. The question of whether SWFs would increase citizen support for or dependence upon exploitation of foreigners and noncitizen immigrants strikes me as especially worth considering. The second piece references an older piece by Doug Henwood, “Pension fund socialism: the illusion that just won’t die”, also worth reading.

Update History:

  • 12-Sep-2018, 1:10 p.m. PDT: Drop unclosed scare quote on “public ownership”; “…if our SWF grew large enough to generate dividends at the scale we might desire…”; “…would not meaningfully be its own, and so become less sympathetic…”; “If the SWF grows towards full-UBI-scale full-UBI scale, it…”; “…the government might well do a shitty job of representing…”; “…ridiculous electoral contests to the accidents of who happens…”; “…can affect the welfare of far-flung and indirectly connected stakeholders…”; add closing semicolon to “&mdash” wannabe entity; “Large, public firms are run by a officers”; “Those who value their own liberty above…”; “…could create political space to enact more distributive redistributive taxation.”; “As it would have if it actually formed the SWF…”; “would ‘naturally’ have occurred had there been an a SWF”; “…concerned about the incredible inequality of wealth of in the contemporary United States (and, indeed, of in most countries…”; “…they will be less efficient purchases are at taking a share…”; “…lending increasingly risky to, and rare among, bank decision-makers.”; “Constituting an a SWF, like…”; “Balancing the task of vigorously vigorous policing of corruption against”; “I think that this sort of compression, but not elimination of variation in economic outcomes…”; “pretty well for the Norway and Alaska, and although. Although you”; “…special cases because of oil revenues, their the inequality-reducing capital accumulation machinery works of SWFs will function regardless…”
  • 13-Sep-2018, 11:50 p.m. PDT: Added bold update with some critical pieces suggested by @BlueIvyRedWorld (Thanks!), and a pointer to the update section from the introductory paragraph.
 
 

8 Responses to “SWF+UBD”

  1. Anon. writes:

    >In a world where the state largely corners the market on it, do capital prices rise as private sector actors compete for the crumbs the solidarity funds leave behind? If so, the share of ownership required to fund a $12K UBI would fall, somewhat.

    You need to take into account that higher prices = lower expected returns, so you’d need to decrease the % disbursed to keep the fund alive.

  2. Steve Roth writes:

    Great as usual. (Though wordy even for you. ;-) )

    Your $1K/per person $75T calc: I don’t think it’s right, net, if the dividends are taxable. Much less, esp if we had an actually goddam progressive tax system.

    The virtue there: it gets vote-garnering salience from everybody actually receiving the payments, while achieving progressivity/means-testing via (less-salient) taxation.

    The crux for me: 35-50% of comprehensive HH income these days is unearned property income, received simply for owning things. I think everybody should share in (some portion of) that work-free income. Would both deliver and BE a massive prosperity increase, and I Q whether there would be any significant macro costs. I think actually, probably, quite the contrary.

    And thanks for the shoutout! I hope lots of people read that far. Cheers.

  3. Brett writes:

    My biggest concern with extending from existing SWF is that they very much have not displaced private capital, especially when you factor in international sources and competition for the Scandinavian countries. We have no idea how a SWF would act or affect the economy if it’s the “only game in town”, and any potential firm trying to raise funds has to take their money if they want to scale above a certain size.

    But who will finance some business in Peoria that could absolutely be profitable in that very local context, but that would not much make sense to the globetrotters at Aspen Ideas?

    I believe this is where Doug Henwood would tell you that most businesses are self-financed, and small starting businesses are probably financed by mortgages/credit card debt/ borrowing from a family member.

  4. Steve Randy Waldman writes:

    Steve R — Just to clarify, the $75T figure comes from solving the equation 0.04 x == 250M * 12K, with 0.04 representing the annual payout of the SWF’s NAV, 250M the current approximate the adult population of the US, and 12K being the annual amount necessary to fund a 1K/month UBI for each of adult.

  5. […] SWF+UBD – interfluidity […]

  6. Global Famine Cannibalism writes:

    The environment could never tolerate that many Middle Caste people who are known to blow wholesale carbon dioxide into our delicate Stratosphere where no rain washes it out and no wind blows it out. Hell, we can hardly tolerate as many poor people as we have, poor people who cut down the forest just so they can plant potatoes. Sure it’s a great financial plan but it is impractical. Practical would be to raise the PCGGP, per capita gross global product by introducing UV, universal vasectomy for every male over the age of 8 years. Such a program would quickly start a religious war that would kill half the population thus increase the PCGGP by 100% without significant damage to our ecology. Deng Xiaoping the Chinese ruler came close to the vasectomies but did operations on women not on men. Such operations are more expensive than vasectomies and a lot more dangerous. Then again women always take the hit.

    What fools these mortals be
    !

  7. […] View Reddit by kludgeocracy – View Supply […]

  8. stone writes:

    I’m in a favour of a UBI but I’m struggling to understand why we’d want a SWF. I think what we need is for big increases in the unaccounted but real wealth of the wider population as JW Mason described in https://jwmason.org/slackwire/wealth-distribution-and-puzzle-of/ . So measures such as tenants’ rights and representation of workers on the boards of companies (and co-ownership by workers) and considerable labour union powers (eg sector wide collective bargaining) and provision of unfunded PAYGO pensions, health care, child care, education, public transport, affordable housing etc etc. Perhaps a land value tax and a wealth tax would help to prevent the concentration of ownership of what remained of marketable asset values after much the wealth of the nation was already in those cryptic public forms.

    My fear is that, if the state got into a colossal bidding war for marketable assets, it could easily be gamed and would predominantly benefit current owners cashing out and management insiders. Some assets such as land can have almost open ended price inflation. A crucial holding of land can be used to extract a rent basically set by the level of parasitic burden that can be endured by the wider population. So even if only some of the land is held by landlords, the price and rent could be such as to nevertheless still extract the full tolerable burden. So I’m not sure that the power of landlords can be tempered by buying them out at a price of their choosing. Their power can be wiped out however with tenants’ rights, right-to-roam and environmental laws, elimination of farm subsidies and a Land Value Tax.

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>