CAMBRIDGE – One of the many things I learned from Milton Friedman is that the true cost of government is its spending, not its taxes. To put it another way, spending is financed either by current taxes or through borrowing, and borrowing amounts to future taxes, which have almost the same impact on economic performance as current taxes.We can apply this reasoning to the United States’ unsustainable fiscal deficit. As is well known, closing this deficit requires less spending or more taxes.The conventional view is that a reasonable, balanced approach entails some of each. But, as Friedman would have argued, the two methods should be considered polar opposites. Less spending means that the government will be smaller. More taxes mean that the government will be larger. Hence, people who favor smaller government (for example, some Republicans) will want the deficit closed entirely by cutting spending, whereas those who favor larger government (for example, President Barack Obama and most Democrats) will want the deficit closed entirely by raising taxes.As the economist Alberto Alesina has found from studies of fiscal stabilization in OECD countries, eliminating fiscal deficits through spending cuts tends to be much better for the economy than eliminating them through tax increases. A natural interpretation is that spending adjustments work better because they promise smaller government, thereby favoring economic growth.For a given size of government, the method of raising tax revenue matters. For example, we can choose how much to collect via a general income tax, a payroll tax, a consumption tax (such as a sales or value-added tax), and so on. We can also choose how much revenue to raise today, rather than in the future (by varying the fiscal deficit).A general principle for an efficient tax system is to collect a given amount of revenue (corresponding in the long run to the government’s spending) in a way that causes as little distortion as possible to the overall economy. Usually, this principle means that marginal tax rates should be similar at different levels of labor income, for various types of consumption, for outlays today versus tomorrow, and so on.From this perspective, a shortcoming of the US individual income-tax system is that marginal tax rates are high at the bottom (because of means testing of welfare programs) and the top (because of the graduated-rate structure). Thus, the government has moved in the wrong direction since 2009, sharply raising marginal tax rates at the bottom (by dramatically increasing transfer programs) and, more recently, at the top (by raising tax rates on the rich).One of the most efficient tax-raising methods is the US payroll tax, for which the marginal tax rate is close to the average rate (because deductions are absent and there is little graduation in the rate structure). Therefore, cutting the payroll tax rate in 2011-2012 and making the rate schedule more graduated (on the Medicare side) were mistakes from the standpoint of efficient taxation.Republicans should consider these ideas when evaluating tax and spending changes in 2013. Going over the "fiscal cliff” would have had the attraction of seriously cutting government spending, although the composition of the cuts – nothing from entitlements and too much from defense – was unattractive. The associated revenue increase was, at least, across the board, rather than the unbalanced hike in marginal tax rates at the top that was enacted.But the most important part of the deal to avert the fiscal cliff was the restoration of the efficient payroll tax. I estimate that the rise by two percentage points in the amount collected from employees corresponds to about $1.4 trillion in revenue over ten years. This serious revenue boost was not counted in standard reports, because the payroll-tax "holiday” for 2011-2012 had always been treated legally as temporary.It is true that some macroeconomic modelers, including the Congressional Budget Office, forecasted that going over the cliff would have caused a recession. But those results come from Keynesian models that always predict that GDP expands when the government gets larger. Entirely absent from these models are the negative effects of more government and uncertainty about how fiscal problems will be resolved.Another recession in the US would not be a great surprise, but it can be attributed to an array of bad government policies and other forces, not to cutting the size of government. Indeed, it is nonsense to think that cuts in government spending should be avoided in the "short run” in order to lower the chance of a recession. If a smaller government is a good idea in the long run (as I believe it is), it is also a good idea in the short run.Robert J. Barro is Professor of Economics at Harvard University and a senior fellow at the Hoover Institution, Stanford University.Copyright Project Syndicate