Before publishing the latest intergenerational report, Joe Hockey, the Australian Treasurer, said that people are “going to fall off their chairs” when they see some of the graphs in it.
I didn’t fall off my chair. The picture of government debt levels painted by the report is not as bad as I was expecting. The report suggests that under current legislation – the budget measures that have obtained Senate approval – net debt would increase to around 50% of GDP by 2055, slightly higher than the current level in Germany and substantially below the current level in the United States and United Kingdom.
The reason why net debt is not projected to deteriorate as rapidly as I had anticipated is bracket creep – the movement of individuals into higher income tax brackets as a result of inflation. Bracket creep is projected to result in a substantial increase in the ratio of tax revenue to GDP over the next five years.
Bracket creep is a sneaky way to raise additional revenue. In the past, governments have adjusted tax rates downwards from time to time to compensate for bracket creep, but there is no provision for that to occur automatically under current legislation.
Consideration of bracket creep helps to focus attention on the main problem with government deficits and growing debt levels – namely, the need for debts to be serviced and repaid from government revenue of one kind or another. Some readers might like to remind me at this stage that future Australian governments could follow the example of the Greeks, among others, and neglect to service debts. The downside of that strategy is that if you ever want to borrow again the cost of funds is likely to be much higher, to reflect the risk of default.
The intergenerational report notes that bracket creep can reduce incentives to work, particularly by people on low incomes. This implies that the marginal excess burden (MEB) associated with bracket creep might be fairly high.
So, what is MEB? In broad terms, taxes impose an excess burden or deadweight cost to the extent that aggregate income is reduced by imposing the tax, for example as a result of reduced incentives to work or invest. (That definition of excess burden is not perfect but it has the virtue of being reasonably easy to understand.) The MEB is the excess burden associated with a small tax increase as a percentage of the additional revenue collected. Thus, for example, a 40 percent excess burden means that for each $1.00 of additional revenue raised, there is an associated deadweight cost of 40 cents. With an MEB of that magnitude, each additional dollar of revenue raised - for infrastructure spending for example - would need to yield a return (discounted to present value terms) of at least $1.40 in order to be worthwhile.
A few years ago the Henry Tax Review commissioned a major study by KPMG (Econtech) of the marginal excess burden (MEB) associated with different forms of taxation in Australia. While the estimates of MEB ranged up to 90 percent, more typical estimates were 40% for taxes on capital income, 24% for labour income, 8% for GST and 2% for municipal rates.
Tax revenue from bracket creep could be expected to involve an MEB somewhat higher than the MEB on labour income. On that basis, the projected increase in tax revenue from bracket creep, equal to about 2% of GDP, could be expected to involve an excess burden of around 0.5% to 0.8% of GDP. That is quite a lot of wealth to be throwing away on top of the deadweight cost of about 6.5% of GDP associated with current levels of taxation.
The obvious way to reduce excess burden is to substitute taxes with relatively low MEBs, such as GST, for income taxes. That would raise concerns that lower income earners might be worse off, but I doubt whether those concerns would be justified if an increase in GST was used to compensate for bracket creep.
The fundamental problem that cannot be overcome by replacing one tax by another to fund high levels of government spending is that MEBs rise exponentially as the tax rate rises. Thus, as shown in the Figure below, if relevant elasticities result in a moderate MEB of 10 percent with a tax rate of 10 percent, the MEB rises to 37.5 percent with a tax rate of 30 percent and to 83.3 percent when the tax rate is 50 percent. I constructed the graph for a report entitled “How Much Government?”, prepared for the New Zealand Roundtable in 2001. The reasoning behind the graph is an application of the standard tools of economic analysis and is spelled out in the report (page 53).
So, what are the implications for excess burden of taxation of the net debt projections in the intergenerational report? The report estimates that the projected net debt of around 50 percent in 2055 under current legislation would involve an increase in net interest payments from 0.7% of GDP to 3.5% of GDP. Under current legislation, assuming an MEB of 40%, there will be an additional deadweight cost of around 1.2% of GDP just to pay interest on government debt. If the government raises taxes sufficiently at that point to eliminate the 6% deficit (underlying cash balance) that would double the additional deadweight cost estimate. Adding on the deadweight cost of bracket creep, gives a total additional deadweight cost of increased taxation equal to about 3% of GDP.
Another way to look at the issues involved is to consider the excess burden of taxation required to pay for the projected increase of government spending equal to 6.5% of GDP - which is the difference between the projected average ratio under the current legislation (31.2%) and the average ratio for the 40 year period 1974-75 to 2013-14 (24.7%). That is the perspective adopted by Henry Ergas in The Australian, March 9, 2015 (article behind paywall). His conclusion is that the required tax increase would cause “a permanent loss of nearly 3 percent of GDP”.