January 2012

Top of This issue Current issue

Negative Payroll Tax: how to reverse our present approach of paying people to be unemployed and of penalising employers

By P.M.Lawrence


This has been adapted, expanded, reformatted and - hopefully - clarified from an email to Jerry Pournelle that he kindly posted on his blog.

Jerry Pournelle occasionally reminds his readers that unemployment benefits work out as paying people to be unemployed, and that regulating firms' working conditions and pay levels works out as penalising employers for employing workers. He usually summarises this by pointing out that if you subsidise something (the former) you get more of it, and if you tax it (the latter) you get less of it. So making employment carry those things is like making a horse pull a cart through a noose around its neck instead of a proper harness.

When you are in a hole, stop digging

Pretty obviously, things would improve if we simply stopped doing these wrong things. It's not so obvious that that's not enough. Unemployment benefits started out from things like the Elizabethan Poor Laws and Bismarck's Welfare State, not simply out of charity but from a hard headed realism that wanted to buy off the social unrest that was already around and growing because there were ever more people without either adequately paid work or personal subsistence resources, a cost that includes higher levels of "crime of necessity" (the technical name for that cost is "Vagrancy Costs"). In Europe, people started getting into that predicament, that "hole", at the end of the Middle Ages, when society, demographics, the economy, and technology reached certain conditions at about the same time. These conditions and their outcomes may be clearer if we use comparisons from the following table, which is itself reduced from a version with more rows and columns to illustrate other things, e.g. slavery:-

Wages and employment for typical societies at various levels of personal resources
Type of society Personal resources Wages norm Employment norm Unemployment norm
Undeveloped society with low levels of resource use, e.g. many countries on first being discovered by modern Europeans. Readily accessible for most people. High, because few people are even interested in working for others at all, and practically none are interested at low wages. The Iron Law of Wages is not in effect. Low, because few people are even interested in working for others at all. Low, because most people have their own subsistence resources and are outside the paid labour pool.
Partly developed society with intermediate levels of resource use, e.g. many countries under colonialism, many European countries in the late Middle Ages, and many modern countries in Latin America, Asia and Africa. Somewhat accessible for many people, but often inadequate (e.g. because of land expropriations or colonial taxation). Low, because most people do need wages, but only as a top up for their other resources, so they can and must bid them down very low. For these people, half a loaf is better than no bread, even if only a whole loaf is enough, as they have some other bread. High, because most people must seek work and can price themselves into it. Low, because most people must seek work and can price themselves into it.
Highly developed society with high levels of resource use, e.g. many western countries in recent history and today. Only accessible for people with enough personal wealth. High, because most people do need wages at levels high enough to survive on wages alone, so they need to hold out for enough. For these people, half a loaf is no better than no bread, if only a whole loaf is enough and they have nothing else. Low, because most people must seek work but not all of them can price themselves into it. High, because most people must seek work but not all of them can price themselves into it.

Notice how there is a pattern to the reasons, a pattern that breaks and shifts and so produces a non-intuitive pattern of high and low norms for wages, employment and unemployment. Notice also that the last two rows co-exist at the same time these days, and global trade makes countries in different rows compete over their wages despite their different circumstances.

Under the hood

Buying off social unrest with unemployment benefits and the like worked, sort of, in the short term, but at the cost of growing the underlying problems for the future - our present. The spread external cost of having the poor around had just been switched for the spread external cost of funding unemployment benefits - even if the accounting in some countries makes it look as if the unemployed are paying for it themselves out of previously accrued payments. But that also means that just getting rid of the things that make the situation worse, the subsidising of unemployment and the burdening of employers that already grew the underlying problems, would just switch back to the external cost of having the poor around - only now at the higher levels that have been allowed to grow. That external cost means that there is something structural that favours unemployment even when the rest of the economy has been stimulated, so that you have to over-stimulate beyond the optimum for employment to pick up, maybe even beyond what is attainable, or you get a jobless recovery because of an underlying mechanism, a mechanism that is operating more strongly all the time.

Economists have actually been looking into this general class of problems - externalities - for about a century, and they have learned a few things. Arthur Cecil Pigou worked out one approach to deal with them, and later Ronald Coase worked out another. Pigou's approach was to use subsidies or taxes precisely in order to get more of what you want and less of what you don't, but in a careful way that actually reduced overall costs. Common sense tells us that taxes and subsidies always make a net burden, whether directly or from the need for funding to be drawn from elsewhere. But the net actually comes from the excess of the cost over the benefit, an excess which comes about because there is a distortion away from the optimum you would have been nearer without intervention. Pigou's insight was that, if there was already a distortion anyway and you pushed the other way with subsidies or taxes instead, you could get nearer the optimum and maybe even hit it if you had enough information (a near miss didn't matter much, because the amount of sub-optimality is a second or higher order function of the "distance" from the optimum, and so typically gets small well before the "distance" gets small - "close enough for government work", in the American phrase). Of course, there is still the transaction cost of churning everything through the government, so Coase's approach of engineering out the externality with property rights is often better, but it may have the hidden catch of yet another material external cost from having to police the property rights. Either way, getting nearer the optimum is always a change from the status quo, and not only does change itself have a cost but also someone's ox is almost bound to get gored - the optimality is an aggregate, so it isn't always an improvement for everyone involved.

Look here and ...

What has all that got to do with the price of fish, i.e. unemployment? Simply that there are both Pigovian and Coasian solutions to it. Since unemployment benefits etc. are already handled through governments, and wages are what Keynes called sticky, Pigovian wage subsidies - that is, wage subsidies that get nearer optimality rather than further away, as they would if they greatly overshot optimality - can be faster acting than the Coasian solutions (which include Distributism to make the resources needed for work into the workers' property and slavery to make the workers into property, so they raise other issues from any coercion involved). That means that subsidy levels have to be set similarly to unemployment benefits or somewhat below, so that their effect matches the spread cost we need to counteract. That gives wage levels at which people still need paid work but, even at the low end of the range, they can afford to work for a wage lower than they need to survive that is a low enough wage to price everybody into work (and to compete with low paid overseas workers, among other things). That means that people only need and get a top up wage, not a living wage, whatever that is. In effect, this moves countries from the third row of the table above to the second row.

After I had done some work of my own in the area (in Australia), I looked around and found that two professional economists had independently come up with something broadly similar: Professor Kim Swales of the University of Strathclyde, along with his colleagues (in the U.K.); and Nobel winner Professor Edmund S. Phelps, McVickar Professor of Political Economy at Columbia University (in the U.S.A.). So I wasn't simply fooling myself that I knew better than professional economists, since some of those had come to the same general idea, albeit using different analysis and pathways to get there.

... Dig here

The big problems with ordinary wage subsidies are that they need huge amounts of funds and - particularly if they go directly to actual and potential employees, as in a Negative Income Tax - that means huge net outgoings while wages slowly and stickily adjust downward enough to price everybody into work (and they don't affect the spread costs properly if you only apply them to a few people within a larger population still facing and generating the older costs, which is partly why the Negative Income Tax pilot studies didn't work as well as was hoped). We three researchers each found the same way around the problem, a Negative Payroll Tax: integrate the wage subsidies with the taxes paid by employers as a tax break of a set amount (or an amount based on a sliding scale, in Professor Phelps's version) per worker per non-overtime hour worked, raising the nominal rate or level of the carrying taxes or of other taxes to compensate. These tax breaks would apply to existing employees and to the self-employed as well as to the newly hired, indefinitely and not just for a limited period - the system is not what some people have sometimes misunderstood it to be, a simple targeted one aimed at helping only the unemployed, and only helping those until they were back in work (doing that would merely encourage employers to switch existing workers for new ones, and then drop those for yet newer ones after a while, and so on).

This integration of the wage subsidies with the tax system means that actual wages paid out don't have to fall even though their net cost to employers does fall, and that no funds actually have to be paid out by the government but rather the pre-tax break gross tax goes up - so the tax breaks are what I term virtual wage subsidies, not actual wage subsidies. Each time someone is retrenched benefits paid out go up but the carrying taxes received go up to match, and the reverse happens whenever someone is hired - budget neutrality for hiring and firing, with each employer facing the true overall costs of hiring and firing, including the spread costs. This boosts employment, but not through a mechanism that uses aggregate stimulus of the economy; some critics have objected that this approach does not create jobs by stimulating the economy, but they are building in the assumption that employment levels are only connected to the levels of economic activity, and so they are leaving out the mechanism involved. (This mechanism actually does end up stimulating the economy, too, but in a secondary way.)

However, this employment boost is only fully cost effective when the set amount provided by the tax break levels matches the costs of unemployment, i.e. the costs of unemployment benefits and the on costs of funding and administering those, etc. If you set the amount lower or higher, you undershoot or overshoot optimality and the overall cost gets higher than it needs to be - though clearly even quite a large overshoot would be some improvement over what we have now, which is effectively a 100% undershoot. (Readers might like to compare this system of tax breaks to the kind of Payroll Tax "holiday" that was just recently tried in the U.S.A. for a limited period, which may be extended - it doesn't amount to the same thing at all, though it's still worth reforming taxes like that since they add to the burdens and faulty employment incentives these tax breaks aim at undoing.)

This bigger gross tax does not mean any short term changes to net tax, apart from oxen getting gored in industries that have already paid for equipment to replace labour, say (the system is revenue neutral in the short term and at least budget neutral after that, since carrying tax revenue only falls in lock step with falls in unemployment benefits as employment improves - which raises revenue from other taxes). However, this does mean some big numbers in the intermediate calculations, which might frighten some people even though they don't correspond to anything real any more than the large displacement of a ship that almost fills a dry dock means that you need that much water in the dry dock to float the ship - Archimedes's paradox. Professor Phelps's version uses the kind of taxes the U.S.A. already has and applies the tax break using a sliding scale, which keeps the numbers small at the expense of needing more administration and policing. Professor Swales's and my version uses the broad based V.A.T./G.S.T. we already have in our countries, though I wouldn't recommend introducing one just to use as a carrying tax - it hurts a lot of other things too.

An automatic stabiliser, a transition, and a newly practical policy option

So far, this article has looked narrowly at employment issues and taken the rest of the economy as given. But you can only apply these virtual wage subsidies up to the size of the economy, because they can't get bigger than the tax base involved; if you try more, it's like pushing on string - everything goes slack. In the ultimate, if the population got too big or the economy got too small, it would be too much for the economy's current capacity and would work through as Malthusian constraints. Now, many people deny that there are any such limits, because they think we can always push them back. But here we are looking at how to structure the economy properly, which means we have to pay attention to those constraints because that is precisely how we can push them back - if we can - just as the maps ships use mark the land even though ships don't go on land, and indeed precisely in order to help them steer clear of it. The two kinds of things we need to achieve are: keeping up the current capacity, which needs investment; and, mobilising that current capacity. Keynes concentrated on the latter, and that is all we will cover in this article as it is the issue confronting us now - the short term one.

Consider what happens to demand in the economy with these employment boosting tax breaks in place. People stay in work even if the economy contracts, though their incomes could still fall as low as the support point, so demand stays higher than under our current system under which unemployment rises - as much higher as is allowed by the excess of wages over what the tax breaks provide. It works out as though they were receiving a Guaranteed Minimum Income (G.M.I.), only set at that support point level which is lower than enough to live on. This is in turn equivalent to everyone being rentiers in relation to part of their income, the reverse of Keynes's idea of the "euthanasia of the rentiers" in that it multiplies them and dilutes their privileges over others by promoting everybody into that group. Because they are so many, the wealth equivalent of that part of their income produces a sizeable Pigou/Real Balance Effect which works as an automatic stabiliser, particularly if the support point is not adjusted for inflation or deflation. (This is not in addition to the gain from more people staying in work - adding that would be double counting - but a different way of describing the same thing that highlights different features, just as different kinds of maps of the same territory do.)

We can go further. There is some latitude to the implementation and administration of this tax break scheme. If it were all handled by issuing anonymous, transferrable vouchers for employees to give to their employers for them to qualify for the tax breaks, the vouchers would soon become money - a "circulating medium of exchange". If the vouchers were then issued by specially endowed funds, say as the dividends on special bonds that formed the endowments, it would all become more at arm's length from the government. Ultimately, once things settled, the endowments could be gradually handed off to families and then to individuals, and then diversified into productive investments, and there would have been a transition to a Coasian approach of a Distributist sort. All that would be needed after that would be something that rebalanced things whenever people's endowments eroded; without that in place before this stage, matters would simply deteriorate as and when that erosion alienated people's subsistence support base. In effect, this moves countries that Negative Payroll Tax put in the second row of the table above again, this time to the first row. But all that really relates to the longer term, and I only describe it to show the essential long run equivalence of the different approaches.

Over and above that, some currently impractical economic stimulus methods may well become newly practical. For instance, it has long been known that you can stimulate an economy without inflation if you raise both taxes and government spending to suitably chosen and often high levels, since although government spending both grows the economy and raises inflation, and taxing both shrinks the economy and lowers inflation or even causes deflation, those pairs of effects don't stand in the same relation to the government spending and taxing that caused them. That means that those pairs of effects don't cancel out at quite the same points when you raise both taxes and government spending, and it turns out that you can find a curve of points where inflation and deflation cancel but the economy is still stimulated. The problem with applying this is political, since individuals can get caught in the middle. However, with this tax break system making its carrying taxes progressive, it should be possible to do all this using those taxes and not taxes that fall directly on individuals. Used with caution, governments could have more room to manoeuvre - though, of course, all that would probably still have an adverse effect on long term investment (not covered in this article), so it should be kept in reserve for emergencies.

What went wrong when we tried this

Well, if this is so clever, why isn't everybody rich? All three of us researchers have tried to get the message across to our respective political establishments, only to be repeatedly listened to politely and then sidelined without being given sound reasons, or indeed any. It's almost enough to make you think there are vested interests in keeping people dependent on a drip feed of the kind of support only they can provide...

References and links to more details

Anyhow, readers might be interested in this for its own sake, and who knows, some aide to Newt Gingrich or to someone like him might pick up on it and pass it on to him. If anybody wants to know more, when I last checked some of Professor Swales's and his colleagues' work was here, some of Professor Phelps's work was here (see also his book Rewarding Work), and I have some here, at a Liberal Party Resolution and following, and at a submission of mine to the Henry Tax Review - but my numbers are a bit out of date by now, e.g. where I mention A$10,000, the figure for late 2011 should be nearer double that (the Australian dollar and the U.S. dollar are close to parity as I write, so there is no great need to adjust the figures as between Australia and the U.S.A.). Professor Swales's modelling indicates that, as my work suggested, there is no overall cost as both GDP and employment levels increase - GDP about half as much as employment levels in percentage terms.

Loose ends

There are some other issues to do with the investment needed to build an economy's current capacity and avoid the Malthusian constraints (though I do touch on them lightly in Appendix C of my Henry Tax Review submission and in other parts of that submission, and in another article of mine), and to do with how one country's economy and tax/subsidy system interfaces with those of others (which I am still working on), but that's a whole other story for another time. Those issues don't matter much right now as they are slower acting than what we've already covered. Suffice it to say that other externalities are at work there, too, so that when a company outsources and incidentally shifts some of the current capacity from one country to another, that's the economic equivalent of a wealth transfer - a giveaway - when one country gets some of another's tax base, particularly but not only that of corporation tax.

The author, P.M.Lawrence, has some other publications here, many on economic issues.