|Saturday, December 13, 2014||(Comment)|
The Georgist dimension of fiscal devaluation: Why it isn't a zero-sum game
Payroll tax, like VAT, is a tax on value added, except that payroll tax exclusively targets the value added by labour, whereas VAT does not discriminate between factors of production. Thus payroll tax and VAT have similar propensities to be passed on in prices.
In an open economy, however, border-adjustment of the VAT (which makes the VAT a consumption tax) becomes significant. An inland payroll tax captures the value added by labour to locally produced goods and services, including those intended for exportation, but not the value added by labour to imported goods and services up to the point of importation. In contrast, an inland VAT captures the value added to locally consumed goods and services, including the value added to imports up to and beyond the point of importation, but not the value added to exports. Hence, if we cut payroll tax and raise VAT, exports become cheaper and imports become dearer. This effect resembles that of a currency devaluation but is achieved by a purely fiscal policy. Hence the policy has become known as fiscal devaluation.
From the name and the explanation for it, one could get the impression that fiscal devaluation is a zero-sum, beggar-thy-neighbour policy. But it isn't. One reason is that the base of the VAT excludes not only exports, but also capital formation within VAT-registered enterprises. Not all countries can be net exporters, but all countries can be net investors in productive capital.
Fiscal devaluation creates jobs by reducing the cost of labour for employers, without reducing workers' take-home wages, and without widening after-tax wage inequalities. The additional incentive to hire labour, due to its lower cost, is offset by the higher consumption tax only if the product of that labour is intended for domestic consumption — not if it is intended for exportation or capital formation. If there were no extra jobs, the tax adjustment would mean higher retail prices for imports but little change in retail prices of local products. The effect of the extra jobs is to allow less welfare spending and more consumption, hence a lower revenue requirement from a larger VAT base, hence a smaller rise in the VAT rate, hence a fall in retail prices of local products and a smaller rise in retail prices of imports. For working people, the effect on imports is a small price to pay for better earning opportunities and cheaper local products. Only those outside the workforce might need additional compensation for the rise in import prices, and only in the unlikely event that it outweighs the fall in prices of local products.
As fiscal devaluation is unequivocally beneficial, one might reasonably wish to implement it on the largest possible scale. Cavallo and Cottani (2010) note that social security taxes in Greece amount to “44% of gross wages, of which 28% is paid by the employer and 16% by the employee.” In order to reduce the cost of labour and stimulate employment, they suggest that the employer-funded component — which they call payroll tax — should be entirely replaced by VAT. They further suggest that payroll tax, rather than being simply abolished, should be offset against VAT (that is, allowed as an input credit). This arrangement would remove payroll tax from the cost of labour but otherwise leave social security unchanged “while rewarding the firms that comply with their payroll tax obligations.”
The trick of offsetting labour taxes against VAT can be applied not only to payroll tax, but also to employee-funded social security contributions and employees' pay-as-you-go (PAYG) personal income tax. In Greece, as in most jurisdictions, “employee-funded” social security contributions are withheld from wages by employers. The same is true of employees' PAYG “personal” income tax. In employers' accounts, withheld social security contributions and “personal” income tax look much like payroll tax; they are additions to the cost of labour, over and above what is paid out to employees. If all imposts withheld from wages by employers were offset against VAT, along with payroll tax, the benefits would be similar in nature, but far greater in degree, than if only payroll tax were offset. Furthermore, because the offsets would encourage employers to pay workers in money rather than fringe benefits, there would be no need for separate taxation of fringe benefits. I have described this proposal as “maximalist” fiscal devaluation (Putland, 2013).
Whenever it is claimed that a switch from personal income tax to VAT would raise prices, including prices of local products, it is assumed that the PAYG “personal” income tax presently withheld by employers would instead be paid out in wages and salaries and would therefore be unavailable to pay the VAT. No such problem arises under “maximalist” fiscal devaluation; the withheld PAYG tax is effectively renamed as VAT before it leaves the hands of employers.
In a country with a floating currency, the benefits of a fiscal devaluation would lead to a currency appreciation, causing a partial rollback of those benefits — but only a partial rollback, because a complete rollback would take away the reason for the currency appreciation. Therefore, although fiscal devaluation was invented as an escape from the constraints of a common currency or fixed exchange rate, it is still worth doing under a floating exchange rate.
So ends my brief introduction to fiscal devaluation. But what has this to do with Henry George (1839–1897) and his “Single Tax” on land values? And how is it that the present writer, a self-confessed Georgist, is not only a defender of fiscal devaluation but also the inventor of its “maximalist” form?
The Georgist paradigm yields three insights concerning fiscal devaluation. First, it offers a further explanation for the benefit of the policy and the limited scale of that benefit. Second, it explains the superiority of the simple form of fiscal devaluation (taxing consumption instead of labour) over a currency devaluation or a more accurate, and consequently more complex, fiscal emulation of currency devaluation. Third, whereas fiscal devaluation uses the offsetting of labour taxes against VAT as compensation for an increase in the VAT rate, the same offsets can serve as compensation for the introduction of a broad-based holding charge on land values, and are better suited for that purpose than cuts in personal income tax.
The first insight concerns the impact of taxation on the three factors of production. Land, as Georgists incessantly point out, is the only factor that can be taxed without reducing production. If we tax labour, there will be less labour. If we tax (produced) capital, there will be less capital. But if we impose a holding tax on the value of land, there will not be less land. Nor will the land be used less productively; someone must own it, and whoever owns it must generate income from it to cover the holding tax, which will not penalize the income so generated. (For present purposes we may take land to mean not only “location, location”, but all assets that taxpayers cannot create or destroy or move, including government-granted privileges.) In particular, owners of vacant or underutilized urban land will be induced to build housing or business accommodation, or to sell the land to someone who will. Thus the tax system will be most conducive to production if it targets only the value of land. A payroll tax or a wage tax targets only labour. A VAT, which targets the price of consumption, is indifferent to whether that price is received as, or paid by, the return to labour, capital or land. This indifference, though less efficient than targeting land alone, is more efficient than targeting labour alone.
The Georgist explanation can perhaps be made more convincing to modern economists by recasting it in terms of the effects of taxes on prices.
Because a land-value tax does not reduce production, it does not raise prices. Not being a cost of production, it does not change the level of production, hence the price, at which a firm maximizes its profit. As the tax is levied on land values, any enterprise that tries to “pass on” the tax in prices will be undercut by a competitor on lower-valued (i.e. lower-taxed) land. A tax that is not passed on in prices is not inflationary and therefore does not add to the so-called natural rate of unemployment, which is the unemployment rate that causes just enough downward pressure on wages to give stable inflation.
(In normal times, central banks adjust monetary policy, usually expressed in terms of an official interest rate, so as to maintain unemployment at the “natural” rate. Sometimes they further obfuscate their agenda by calling the “natural” rate the “full-employment” rate. But whatever they call it, they have not the faintest intention of allowing everyone who wants a job to find one.)
In contrast to a land-value tax, a VAT notoriously raises prices and therefore raises the “natural” rate of unemployment. But the damage is moderated in two ways. First, the effect on prices does not extend to exports and capital formation; in other words, a VAT raises consumer prices but not producer prices. Second, a VAT incidentally captures price markups that are attributable not to productive effort, but to protection from competition, including the protection conferred by superior location. In such cases a VAT does not add to prices, but subtracts from the economic rents that pad existing prices.
A labour tax has fewer redeeming features. It raises not only consumer prices but also producer prices. To the extent that it reduces economic rents, it does so not by scaling down the markup but by scaling up the cost base, for which there are no input credits, hence no hedges against higher costs, hence a higher risk that economic rents will give way to losses. By concentrating exclusively on the labour component of the cost base, it maximizes the magnification of labour costs, hence the inflationary effect of any increase in labour costs due to lower unemployment; thus, for a given amount of revenue, a labour tax maximizes the increase in the “natural” rate of unemployment.
Whether the effects of taxes on employment are direct or mediated through inflation, the implication is the same: fiscal devaluation reduces unemployment, but the replacement of labour taxes by a land-value tax would reduce it further.
The second insight arises from the further tax adjustments that are needed to turn fiscal devaluation into a more accurate emulation of currency devaluation. According to Farhi, Gopinath & Itskhoki (2013), these adjustments include an increase in taxation of personal income from labour, and a reduction of “capital taxes to firms”. By Georgist criteria, the former is destructive, while the latter is constructive only in so far as “capital taxes” fall on produced capital, not land — a distinction that the authors fail to make. The authors explain that the reduction in “capital taxes” avoids “an incentive to substitute labor for capital, an effect absent under a nominal devaluation.” As central bankers try in vain to reduce unemployment to levels that even they consider acceptable, one could be forgiven for asking whether an incentive to substitute labour for “capital” is altogether evil, especially when the labour so favoured includes that expended in the formation of real capital, while the so-called “capital” displaced by labour might actually be land. Proponents of fiscal devaluation therefore do well when they choose simplicity over precise emulation of currency devaluation.
The third insight offers Georgists a new solution to an old problem. The “Single Tax” on land values is open to criticism on the ground that retirees, who are still likely to live on valuable land but no longer likely to pay large amounts of personal income tax, would not be adequately compensated by the abolition of existing taxes. This objection assumes that PAYG personal income tax would be abolished in the hands of employees. If, instead, the PAYG “personal” income tax withheld by employers were offset against VAT without increasing the VAT rate, there would be a general fall in prices, the benefits of which would extend to retirees. Thus, for the purpose of compensating for the new impost on land values, the offsetting of PAYG “personal” income tax in the hands of employers is preferable to its abolition in the hands of employees. For land owner-occupants with limited income, a further concession has long been embraced by Georgists: any unpaid portion of the land-value tax can be allowed to accumulate as a debt against the land, until the land title is transferred in the normal course of events, and the debt can be capped to the value of the land (not including any buildings).
Under a “maximalist” fiscal devaluation, the downward effects on prices, due to the offsetting of labour taxes paid or withheld by employers, compensate for the upward effect due to the increase in the VAT rate. If, instead, there is no increase in the VAT rate, there is a net reduction in prices, which in turn can compensate for a new charge on land values. It is even possible to set the VAT rate to zero, causing a larger fall in prices, compensating for a higher charge on land values, which replaces a larger amount of forgone revenue.
(Of course, if there are offsets against a zero-rated VAT, the revenue from VAT is negative. This is not as strange as it may seem. Under a conventional VAT, a newly established business may have a temporarily negative VAT bill, because it may reclaim more VAT on its set-up costs than it collects on its early sales, while an export-oriented business may have a permanently negative VAT bill, because it does not collect VAT on its export sales but still reclaims VAT on its domestic expenses. Under a zero-rated VAT with offsets, negative VAT revenue becomes the norm. And because the effect of VAT on prices is better known than that of labour taxes, negative VAT revenue helps to frame the message that offsetting labour taxes against VAT reduces prices.)
If a holding charge on land, payable by land owners, replaces taxes that feed into prices and raise the “natural” rate of unemployment, the owners will be compensated by lower prices and better job opportunities. For residential tenants, the benefits will be only partly competed away in higher rents, because the increased supply of housing will improve tenants' bargaining power. Households will therefore have more capacity to consume, so that businesses will enjoy higher sales. Hence the beneficiaries will include commercial tenants; and again, due to the increased supply of accommodation, the benefit will be only partly competed away in higher rents. For residential and commercial landlords, the benefit of tenants' higher capacity to pay rent will come in addition to the benefits of the reduction in the cost of living, further compensating the landlords for the holding charge on their land.
The transition to the new system will not be traumatic. Anticipation of the land-value tax will cause a burst of construction as land owners prepare to earn income to cover future holding costs. Anticipation of lower prices for goods and services will cause a hiatus in consumption, releasing productive capacity for the burst of construction.
The first stage of implementation is to set the VAT rate to zero, offset all labour taxes paid or withheld by employers against the zero-rated VAT, abolish separate taxation of fringe benefits (because there is no longer any incentive to pay employees in kind), introduce a heavy broad-based land-value tax, and abolish personal income tax on rental income from property (which is made redundant by the land-value tax).
The first stage eliminates most of the compliance cost of the VAT; if the rate is zero, there is no need to calculate the base. But the compliance costs of the labour taxes remain, although these taxes, being offset against the VAT, no longer raise any revenue.
Hence the second stage of implementation is to get rid of the revenueless taxes and, in so doing, to get rid of the associated offsets and compliance costs. The saving in compliance costs leads to a slight reduction in prices.
Different labour taxes require different mechanisms for their abolition. The simplest case is a payroll tax used for general revenue: the tax is abolished and the offset consequently disappears, with no change in revenue. An off-budget employer-funded social security contribution, such as Australia's Superannuation Guarantee, is only slightly more complex: employers no longer pay the contribution, and the government, instead of reimbursing the employers through an offset, funds the contributions directly from general revenue. The same applies to “employee-funded” social security contributions withheld by employers, except that employment contracts and awards must be reinterpreted so as to preserve wages and salaries net of social security contributions, ensuring that the abolition of the contributions is a saving for employers, leading to lower prices. Similarly, if the abolition of PAYG “personal” income tax is to be passed on as a reduction in prices, employment contracts and awards must be reinterpreted so as to preserve wages and salaries net of “personal” income tax. In the last case, as in the first, there is no interaction with social security contributions; the tax is abolished and the offset consequently disappears, with no change in revenue.
The above mechanisms of abolition are equally applicable to a fiscal devaluation with a positive-rated VAT and to a Georgist reform with a zero-rated VAT. But in the latter case, when all the taxes offset against the zero-rated VAT have been abolished, the VAT raises zero revenue and can itself be abolished.
In summary, Georgists and proponents of fiscal devaluation can learn from each other: Georgist theory helps to explain the positive-sum benefits of fiscal devaluation, while the tax offsets associated with a “maximalist” fiscal devaluation can serve as components of an ambitious Georgist program.
Farhi, E., Gopinath, G., & Itskhoki, O., “Fiscal devaluations” (draft), 6 April 2013; www.is.gd/FarhiGopinathItskhoki2013 (PDF).
Putland, G.R., “Fiscal devaluation on steroids”, Macro Business, 3 September 2013; www.is.gd/Putland2013. Abridged as “Maximalist ‘fiscal devaluations’ for Greece and Australia”, grputland.com, 5 October 2013; www.is.gd/maxifd.
[Drafted 6 December 2013; posted 13 December 2014.]
|Tweet||Return to Contents|