Saturday, July 2, 2016

The Universal Basic Share

Universal basic income, or UBI for those acronymically minded, is in the news these days, along with other brilliant post-modern inventions such as Brexit or Trump. Unlike these other luminaries, though, UBI is a genuinely cool idea: give everyone a basic amount to spend, and let them do what they will with it. They could write poetry, compose sonatas, or study number theory. They could work for more income if they wanted. Or they could relax and do absolutely nothing.

UBI is the offspring of a beautiful dream: the liberation of the human being from the drudgery of everyday labor. But it is also the product of a scary thought: the trend of ever-advancing automation, now accelerated many-fold by new deep learning algorithms.

You see the connection, of course. If a bunch of creepy robots are going to pass the Turing test at the call center, or drive up shinily when you hail an Über, or stack boxes even as they are energetically prevented from doing so, or even dance while doing the dishes, you'd better find something better to do with your labor time.

So UBI is a nice gesture; it sends you on your way with a little stash of income that you can do with as you please. But of course, a little stash multiplied by the population ends up -- not surprisingly -- in a big stash. Try giving everyone in the United States $10,000 each annually, and you will see that the required payout comes to a cool 3 trillion per year, which is in excess of three-quarters the annual federal budget. Yikes!

Nevertheless, the idea has found serious purchase in Europe: the Swiss even voted in June on UBI of around $2500 per month per adult. It didn't pass --- in fact it was turned down by a large margin -- but a serious warning shot had been fired. Finland and the Netherlands are planning to trial UBI by following a group of lucky recipients around and seeing what they do with their monthly payments. (Though somehow the thought of being stalked by a group of randomistas asking how I am spending my UBI is weird; I know what I'd tell them.)

You would think that the UBI is a good idea for rich countries. But there is also a prima facie case for trying it in a country like India, which one way or the other has been making very large transfers for decades.  Just the public distribution scheme for foodgrain represents a subsidy of around 1.4% of GDP, but if you add to this the subsidies on fertilizer, transportation, water, electricity and other goods, we are up to well over 4% of GDP. Then there are the so-called "revenues foregone" through various exemptions, chiefly via relief on excise and customs duty, that will take you into the region of another 6% of GDP. We're now up to 10% and counting, and we're counting because these are just in the domain of the Central Government; there are more subsidies at the State level, and there are other implicit subsidies via sub-market pricing of public sector goods. (See also Santosh Dash's comment below.)  I'm not counting large sources of social expenditure, such as education and health, nor the national rural employment guarantee scheme,  which provides every rural household the right to 100 days of work at a basic wage. Here's an illuminating note on central Indian subsidies put together by Siddharth Hari, a doctoral candidate at NYU.

These subsidies are often greatly lamented, largely on the right, by individuals who blame them for all sorts of bad outcomes. One favorite lament is that there are big leakages due to corruption. Another is that subsidies are often mis-targeted (over and above the corruption) to the non-poor. And the libertarian spirit typically completes this tri-headed litany: why should the Government tell us what to eat, or how many health checkups to have? And what is it doing in the food distribution or transportation business, or in any business for that matter? Why not just hand out plain unvarnished -- and presumably untarnished -- cash instead to everyone, and be done with it?

I want to refrain from engaging in that debate here, but the bottom line is this: talk of a universal cash transfer that replaces a system of multifarious, nefarious transfers has long been in the Indian air.  So it comes as no particular surprise to learn that careful, long-standing observers of the Indian economy have promptly added two and two to ask: can we cobble together a basic, unconditional, universal income for all of India's citizens?

You might justifiably and indignantly ask: unconditional and universal? Why should the rich also be treated to free income? Answer: try targeting, and the leaky bucket will emerge again, spilling copiously. But isn't a universal transfer the logical equivalent of a bucket with basically no bottom at all? Perhaps, but then we'd spend all our time issuing and examining BPL cards, and given the massive corruption and incompetence in the bureaucracy, you may as give everyone the money and save us the headache. Well, ok, but I don't feel like providing Ambani with an assured UBI. Oh, we can get around a lot of that by requiring that the claimant must show up in person bearing an identity card to claim her income. Ambani won't show up in person. A lot of the rich won't show up. But who's to say that the bureaucrat won't claim that they did show up? And so on and so forth. Or you could attack all of this from another direction: won't the poor squander their cash cavorting and drinking, as the poor apparently do? Then the usual arguments about paternalism can start up. There is no end to this.

In this post, I am going to tentatively accept the idea of universality and non-paternalism, and look at the other elephants in the room (alas, there's a veritable herd of them):

1. The promise of an UBI can be inflated away. Who's going to make sure this thing is properly indexed to rising prices, and what if it's not? An unsympathetic Government can erode all the promises --- all the subsidies and the transfers that were so clumsily but irrevocably made in kind --- and make them vanish into thin air in a matter of years. (With inflation at 5%, a nominal commitment in fixed rupees with halve -- in real value -- in 14 years.)

2. The commitment looks really huge (sorry, I seem to have inadvertently quoted Trump).   In 2014, the Rangarajan Committee submitted its report proposing a monthly poverty line of Rs. 972 and Rs. 1407 (urban). With rural population shares taken into account, that's a bit north of Rs. 13,000 ($200) per year per person. A pittance? Yes. But multiply by India's population of 1.25 billion and you're at around 12% of India's GDP ($2.09 trillion in 2015).  If you want to cut that back to Rs 10,000 per year (or around $150), you're at 9% of GDP. So there you have it, ladies and gentlemen: 9-12% of GDP to bring every man, woman and child up to speed, or at least walking pace.

Is this do-able? It all depends on whether those huge subsidies to the non-poor can be removed. Pranab Bardhan writes:

"[T]he Indian government doles out significantly more than [10% of GDP] in implicit or explicit subsidies to better-off sections of the population, not to mention tax exemptions to the corporate sector. By discontinuing some or all of these subsidies – which, of course, do not include expenditures in areas like health, education, nutrition, rural and urban development programs, and environmental protection – the government could secure the funds to offer everyone, rich and poor, a reasonable basic income."

There's some more optimism expressed by Abhijit Banerjee and by Guy Standing, but the political economy of subsidy removal does look menacing, to say the least. Central government expenditure as a share of GDP has been declining since 2010; this year it will be a bit more than 13%. That matches the demands that UBI would make, which isn't comforting at all. Nor is it comforting that no one pays taxes in India. In a more pessimistic piece, Maitreesh Ghatak concludes that:

"A universal cash transfer scheme is therefore not feasible without raising additional taxes. Not just that, given that only 1 per cent of Indians actually pay income tax, while a mere 2.3 per cent file tax returns, the fiscal instruments to claw back the transfer from the rich do not exist."

It does seem like we're on a dramatic edge here, and a lot must hand on whether existing subsidies can be credibly removed.

3. For my last elephant, let's go back for a moment to this whole automation business. Some years ago, I observed in this post  (a tad gloomily) that:

"to avoid the ever widening capital-labor inequality as we lurch towards an automated world, all its inhabitants must ultimately own shares of physical capital. Whether this can successfully happen or not is an open question. I am pessimistic, but the deepest of all long-run policy implications lies in pondering this question."

I've italicized the phrase I want to emphasize here: if we're truly headed towards automation, it is not enough to pay out UBI and let a small group of residual claimants eagerly divvy up the remaining surplus. Even with indexation to inflation, the UBI is a fixed commitment. What happens, then, as profits continue to rise in business? Is no share to be passed on to the population? Will class warfare be reduced to annual debates about how to adjust the UBI?

In the rest of this article, I'm going to propose a simple amendment of UBI that holds out serious hope for dealing with all of these issues and more. I'm going to call it the universal basic share, or UBS. Simply put, the UBS is a commitment that is expressed, not as a sum of money, but as a share. Specifically, I propose that we commit a fixed fraction of our GDP to the provision of a universal income for all.

Consider six merits of this proposal, not necessarily in order of importance:

A. It is country-neutral. It can be introduced into every country, rich or poor. It scales up or down with country-level income.

B. We can start small.  In the Indian example, the numbers do not have to be at Rs. 10,000 to begin with. But over time, they will get there. In this sense, the proposal takes (some) care of the debate that we "cannot afford it."

C. The UBI commits a government to pay out a fixed sum, come hell or high water. In contrast, UBS insulates against shocks to the fiscal system that are correlated with GDP shocks.  (Given the amounts involved, one might imagine even rich governments being risk-averse.) But the upside to the general public will be enormous.

D. The UBS does not need to be indexed at all. It's fixed as a share of nominal GDP,  and that will automatically take care of any indexing that's needed.

[Update 1: A UBI can be indexed in India using the dearness allowance, which is a cost-of-living adjustment based on the cost-of-living index and paid out to public-sector employees and pensioners. Maybe, though in countries where inflation statistics are dodgy I'd be wary of this. I'd be wary of formula manipulations in India as well, once a truly enormous commitment such as UBI is on the table. In any case, I am after more than mere indexation; see point F below.]

E. The UBS will create an incentive for a majority to demand a better tax collection and auditing system. And the government, too, would be incentivized to close off its tax loopholes. For India, this is a first-order issue.

F. The UBS allows everyone to share in the prosperity of a country.  To me, this aspect of equity-sharing is --- in the longer run --- the most important feature of the UBS. It is our protection against unbounded inequality as we move into an increasingly automated universe.

To implement a UBS, the most important thing is to get the share right. Giving everyone Rs. 10,000 per year takes us to about 9% of GDP. But it's not enough to leave it there; we need a sense of what this looks like as a fraction of government expenditure. This is an extremely tricky business. Let me illustrate with India, which --- given its existing slew of explicit and implicit subsidies --- is possibly one of the most difficult examples out there. (Fair warning:  I have the back of an envelope out as I speak, so the numbers below would need to be refined.) 

The central government's expenditure share as a percentage of GDP is a bit shy of 14% in 2014-2015. But central and state expenditure combined is double that: around 27% in 2014-2015 (here for the gory details). For revenue foregone and other implicit subsidies, which we would need to take back, add on another 6-10%. That gets us to about 35%. So to access 9% of GDP as UBS, we would need to contribute 25% of government expenditure, inclusive of all subsidies, to the cause.

[Update 2: an alternative is to commit UBS directly as a share of government expenditure, which is the form in which I originally suggested it. The linking of basic income to overall prosperity then is less direct. Moreover, as Pranab Bardhan, Karna Basu and Siddharth Hari have pointed out to me,  the government could suffer from possible disincentives in raising expenditure, fearing that part of the increase would be "taxed off" by the UBS. Though in view of the deficit, such fears might be a blessing in disguise. On the other hand, the government would gain better insurance: if there is a sudden fiscal crisis, even one that's independent of a GDP shock, its commitment to UBS would adjust accordingly.]

Can we really usher in the right to a UBS? I have no clue whether we have the political will to pull something like this off. But remember: it's a share that's being committed. At Indian rates of growth and with an improving fiscal system, we can get the resulting numbers to double in 10-12 years, and double again a decade after that. So if we want to start smaller, we can entertain that thought.

Some postscripts on the UBS:

(i) If you want to institute a share, do it when you start the program. Once a number is fixed, no one wants to move towards a share as it looks risky. With a share to begin with --- where there was nothing before --- matters can be very different.

(ii)  For each year, the payout assessment will need to be done. This can be done using the previous year's GDP (or expenditure, in case the variant is tried) and dividing by population estimates. Uncollected payouts --- and hopefully there will be a lot of those --- can go into an insurance endowment or otherwise used.

(iii) [Update 3.] After I wrote this, Rajiv Sethi pointed me to Robert Shiller's proposal to issue trills, which is a government-issued security that would pay a share -- in trilllionths, hence "trill" --- of GDP. Yes! A UBS is certainly viewable as a variant of a gigantic, collectively held trill --- a plain bill, then, perhaps? Look here for a related proposal by Rajiv to hold individual bank accounts at the Fed. In keeping with the adage that there's nothing new under the sun, Ugo Colombino pointed out the connection to the citizen's dividend, which is a form of UBS based on natural resources; Alaska implements a form of this as the Alaska Permanent Fund. In the words of Thomas Paine, "men did not make the earth." Rahul Basu told me about the efforts --- inspired by Alaska's fund--- to secure a permanent fund in Goa. Read here about such a movement, and read here about the Supreme Court directive to create a Goa Iron Ore Permanent Fund.

Hey Switzerland, want to try again?

Thanks: Pranab Bardhan, Karna Basu, Rahul Basu, Ugo Colombino, Parikshit Ghosh, Siddharth Hari, Aditya Kuvalekar, and  Rajiv Sethi.