The Economy Under the President’s Budget and Under CBO’s Baseline Policy Assumptions

In addition to estimating the direct budgetary impact of the President’s proposals (see Chapter 1), the Congressional Budget Office has analyzed how those policies would affect the economy as a whole. CBO estimates that the direct budgetary effect of the President’s proposals would be an increase in the cumulative deficit, relative to that shown in the current-law baseline, of $336 billion, or about 0.4 percent of cumulative gross domestic product, from 2009 to 2013, and a reduction in the cumulative surplus of $651 billion, or 0.6 percent of cumulative GDP, from 2014 to 2018. (Those estimates do not consider the influence of economic feedback.) Taken together, several major provisions would tend to have mutually offsetting effects on the economy: Some that would tend to reduce output would offset other provisions that would tend to expand it, resulting in modest effects overall.

CBO’s estimates of the economic feedback associated with the President’s proposals depend on a variety of specific assumptions. However, under any of the assumptions incorporated into this analysis, economic feedback would modify the budgetary effects of the proposals: From 2009 to 2013, the feedback could raise the proposals’ cumulative impact to about $410 billion or reduce it to about $185 billion. From 2014 to 2018, the feedback could raise the proposals’ cumulative impact to about $670 billion or reduce it to about $460 billion (see Figure 2-1 and Table 2-1).

Notes: The estimates in the panels above reflect the supply-side effects of the President’s proposals on the economy but exclude demand-side economic impact, as explained in the text. A negative change indicates an increase in the cumulative deficit or a decrease in the cumulative surplus relative to CBO’s baseline.

CBO’s analysis used the following models (which are described in the text): (A) "textbook" high model, (B) "textbook" low model, (C) closed-economy life-cycle model with lower government spending after 2018, (D) closed-economy life-cycle model with higher taxes after 2018, (E) open-economy life-cycle model with lower government spending after 2018, (F) open-economy life-cycle model with higher taxes after 2018, (G) infinite-horizon model with lower government spending after 2018, (H) infinite-horizon model with higher taxes after 2018, (I) Macroeconomic Advisers’ model, and (J) Global Insight’s model.

		2009 to 2013	2014 to 2018


Growth Models
Without Forward-Looking Behavior
Textbook Model
	High (Hours worked respond strongly to tax-rate changes)	-328	-655
	Low (Hours worked respond weakly to tax-rate changes)	-349	-672

With Forward-Looking Behavior
Closed-Economy Life-Cycle Model
	Government spending adjusted after 2018	-265	-501
	Taxes adjusted after 2018	-272	-535
Open-Economy Life-Cycle Model
	Government spending adjusted after 2018	-185	-458
	Taxes adjusted after 2018	-200	-496
Infinite-Horizon Model
	Government spending adjusted after 2018	-300	-579
	Taxes adjusted after 2018	-300	-583

Macroeconometric Models
Supply-Side Contribution
Macroeconomic Advisers’ Model		-365	n.a.
Global Insight’s Model		-335	n.a.

Supply-Side and Demand-Side Contributions
Macroeconomic Advisers’ Model		-412	n.a.
Global Insight’s Model		-373	n.a.

Memorandum:
CBO’s Estimate of the Budgetary Effects of the President’s Proposals Under Baseline Economic Assumptions		-336	-651

Notes: A negative number indicates an increase in the cumulative deficit or a decrease in the cumulative surplus relative to CBO’s baseline.

The "textbook" growth model is an enhanced version of a model developed by Robert Solow. The life-cycle growth model, developed by CBO, is an overlapping-generations general-equilibrium model. The infinite-horizon growth model is an enhanced version of a model first developed by Frank Ramsey. The models by Macroeconomic Advisers and Global Insight, which are available commercially, are designed to forecast short-term economic developments. The various models reflect a wide range of assumptions about the extent to which people are forward-looking in their behavior: In the textbook model and those by Macroeconomic Advisers and Global Insight, people have the least foresight, whereas in the infinite-horizon model, people’s foresight is perfect and extends indefinitely to include a full consideration of effects on descendants.

In models with forward-looking behavior, CBO had to make assumptions about how the President’s budget would be financed after 2018. CBO chose two alternatives—adjusting government purchases of goods and services and transfer payments or adjusting marginal tax rates.

Over the long run, the nation’s potential to produce goods and services depends on the size and quality of the labor force, on the stock of productive capital (such as factories, vehicles, and computers), and on the nation’s technological expertise. Changes in those determinants of potential output—which economists call "supply-side" changes—can have a lasting, sustainable influence on the economy’s ability to supply goods and services.

In the short run, however, economic activity can deviate from its potential level in response to changes in aggregate demand. For example, CBO estimates that output is currently below its potential level, in part because of the collapse of the housing market and the resulting turmoil in the financial markets. Such "demand-side" variations can alter the employment of labor and the use of capital relative to their long-term potential levels.¹ Unlike movements on the supply side of the economy, however, changes in demand generally balance out over time: Long-term corrective forces tend to move the economy back toward its sustainable potential determined by the supply side. Nevertheless, when aggregate demand is low (as it appears to be currently), the positive demand-side effects of government policies such as tax cuts or spending increases can hasten a return to potential output.

The government’s budgetary policies can influence the economy through various channels, some of them affecting the supply side of the economy, some affecting the demand side, and some affecting both. Changes in tax rates can affect the willingness of people to work and to save, potentially influencing short-run demand but also affecting sustainable, long-term supplies of labor and capital. Similarly, changes in government spending for goods and services and in government transfers can affect short-run demand and increase or reduce the amount of resources available for private investment, thus affecting the long-term size of the capital stock. (CBO’s macroeconomic analysis distinguishes between the government’s transfers of funds to people and its purchases of goods and services, but most of CBO’s models do not differentiate more finely among the various types of government spending to distinguish specific effects on long-term economic performance. In reality, the economic effects of different types of government spending vary.)

The economic effects of changes in revenues and in spending depend on how those changes are financed. In the short run, reductions in taxes or increases in spending can be absorbed into larger budget deficits. Over the long term, however, other policy changes are needed to offset the loss of revenues or the increase in spending and prevent unchecked growth in government debt relative to output.² Those policy changes significantly influence the long-term economic effects of the initial change in spending or revenues.

The President’s budgetary proposals can affect the quantity and the quality of the labor force, the size and composition of the capital stock, and the strength of the nation’s technological progress. Each of those supply-side effects helps determine the course of potential economic output.

Labor Force. Potential output is strongly tied to the supply and the quality of labor in the economy. A sustained, long-term increase in total hours worked improves the economy’s potential to generate output. CBO’s analysis focused on channels through which the President’s proposals could affect the number of hours of labor supplied.

The President’s proposals could affect the quantity of labor through two main channels. First, several of the policies proposed would change people’s after-tax income, although those proposals would not significantly alter the marginal tax rates on income resulting from labor.³ (The extension of the child tax credit would raise the after-tax income of some workers, for example, but it would not affect their marginal tax rates.) In the absence of a change in marginal rates, an increase in after-tax income tends to reduce the number of hours of labor supplied because people can maintain their standard of living with less work; conversely, a decline in income tends to increase hours supplied.

Second, some provisions would change after-tax income and after-tax compensation for each additional hour of work. (The extension of the marginal tax rates on income enacted in the Economic Growth and Tax Relief Reconciliation Act of 2001 would increase both. Other provisions, such as the President’s health proposal, would reduce them in most years.) Provisions that raise after-tax income and incremental after-tax compensation (and provisions that reduce both) have opposing effects on people’s incentives. In the case of the extension of EGTRRA, for example, workers would be encouraged to work longer hours if they earned more for each extra hour of labor they supplied. But a disincentive also is possible: Workers might find they could maintain current after-tax income without working extra hours, so they would consider not increasing their work hours.

On balance, those opposing incentives largely offset one another, although the first set (more pay for more hours) slightly outweighs the second. Reductions in marginal tax rates will tend to increase modestly the hours of labor that workers supply, primarily because those reductions will draw secondary earners (for example, the spouse of a household’s primary breadwinner) into the labor force.⁴ Conversely, increases in marginal tax rates will modestly decrease hours worked.

CBO estimates that if enacted, the President’s policies would lower the marginal tax rate on labor by more than 3 percent in 2008, compared with the rate shown in CBO’s baseline, largely as a result of the proposal to extend for one year an increase in the exemptions allowed under the alternative minimum tax (see Table 2-2).

			Difference
Calendar Year	Tax Rate Under Current Law	Tax Rate Under the President’s Budget	Percentage Points	Percent


2008	30.5	29.4	-1.1	-3.5
2009	30.7	30.6	-0.1	-0.4
2010	31.0	31.0	0.1	0.2
2011	32.3	31.0	-1.2	-3.8

2012	32.5	31.5	-1.0	-2.9
2013	32.5	31.9	-0.7	-2.0
2014	32.6	32.2	-0.4	-1.3
2015	32.7	32.5	-0.2	-0.6

2016	32.7	32.7	*	0.1
2017	32.8	33.1	0.3	0.8
2018	32.8	33.3	0.5	1.5

Note: The effective federal marginal tax rate on income from labor is the share of the last dollar of such income taken by federal individual income and payroll taxes.

The President’s policies would leave marginal rates nearly unchanged in 2009 and 2010. After 2010, two sets of proposals would have opposing effects on marginal tax rates. Proposals that would make some provisions of the 2001 and 2003 tax legislation permanent would reduce marginal rates by nearly 6 percent in 2011 and by somewhat less in the following years (because the AMT would offset an increasing share of the marginal rate reductions over time). However, beginning in 2009, proposed changes in the tax treatment of health insurance premiums would raise marginal rates on labor by gradually increasing amounts.^⁵ Together, those proposals would reduce the effective marginal federal tax rate on labor income by about 4 percent in 2011. As the effects of the health proposals grew larger each year, however, that reduction would shrink, until the net result would be a small increase that began in 2016 and rose to about 1.5 percent by 2018.

Taking into account the effects described above, CBO estimates that the President’s budgetary proposals would probably increase the number of hours people work in 2008 but have only a small effect in 2009 and 2010. In the first few years after 2010, reductions in marginal tax rates might lead workers to increase their work effort. However, over the 2014–2018 period, the President’s proposals would tend to have a modest influence, on average, on the number of hours supplied.

Improvements in the amount of education, training, and experience workers have and in their efforts on the job—all improve the quality of each hour worked—also could result in higher potential output. However, CBO did not incorporate such additional labor quality effects into this analysis because they would not produce a measurable effect in the 10-year projection period.

Capital Stock. The President’s budgetary policies would influence the size of the nation’s capital stock primarily by affecting national saving. National saving is private saving minus the budget deficit; an increase in private saving raises national saving, whereas an increase in the budget deficit reduces it. An overall decline in national saving reduces the capital stock owned by U.S. citizens over time, either through a decrease in domestic investment or through an increase in net borrowing from abroad, or both.

In every year from 2008 to 2018, except 2010, the President’s proposals would expand the federal deficit or shrink the surplus relative to that in CBO’s baseline. However, two proposals that contribute significantly to that outcome would, in CBO’s judgment, have little effect on national saving. First, the establishment of individual accounts in Social Security would expand private saving but add to the budget deficit; the net effect on national saving would be much smaller than would be the increase in the deficit itself. (The net effect of that proposal on national saving depends on the reaction of households to the funds deposited into individual accounts and on the effect of the deposits on federal budget policy.) Second, extending the repeal of the estate tax would produce an increase in income that would nearly all be saved, almost entirely offsetting the resulting decline in government revenues. As a result, the higher deficits that stemmed from those proposals, by themselves, would have only a slight, negative effect on national saving.

Several of the President’s tax proposals might spur private saving by reducing the effective marginal tax rates on capital income and thus raising the after-tax rate of return on savings. Through that channel, the tax proposals would influence private saving in two opposing ways, just as lowering the marginal tax rate on labor income would have opposing effects on the supply of labor: Higher after-tax returns would tend to increase saving and thus reduce current consumer spending, but they also would boost the value of existing assets, making households wealthier and thus tending to encourage spending. On balance, the combined effect on spending of higher after-tax returns can be either positive or negative, and researchers generally conclude that the effect is small. Nevertheless, to cover other possibilities, CBO included in its analysis a range of plausible assumptions about how households might respond to changes in the after-tax rate of return on savings. At one end of the range, some of CBO’s models assumed that the rate would have little or no effect on how households allocated income between spending and saving; at the other end, some models assumed that raising the rate of return would boost saving and reduce spending significantly.

The provisions of the President’s budget that could affect the after-tax rate of return on capital include extending EGTRRA’s marginal income tax rates, extending the cuts in tax rates on dividends and capital gains (enacted in the Jobs and Growth Tax Relief Reconciliation Act of 2003), and expanding tax-free savings accounts. (Appendix B provides greater detail on the potential economic effects of the President’s proposals for dividend and capital gains taxation, tax-free savings accounts, the estate tax, and individual accounts in Social Security.) CBO summarized the effects of most of those provisions on the rate of return on savings by estimating the aggregate effect they would have on the average effective marginal tax rate on capital income, compared with CBO’s estimate of that rate under current law (see Table 2-3).⁶ According to CBO’s estimates, the effective marginal federal tax rate on capital income during the 2011–2018 period would be about 14 percent to 15 percent lower under the President’s proposals than under the policies assumed in CBO’s baseline.⁷

			Difference
Calendar Year	Tax Rate Under Current Law	Tax Rate Under the President’s Budget	Percentage Points	Percent


2008	14.0	14.2	0.2	1.2
2009	14.1	14.1	*	0.3
2010	14.0	14.0	*	0.3
2011	16.3	13.8	-2.5	-15.1

2012	16.3	13.8	-2.5	-15.2
2013	16.2	13.8	-2.4	-15.0
2014	16.1	13.7	-2.4	-14.7
2015	16.1	13.7	-2.3	-14.4

2016	16.0	13.7	-2.3	-14.6
2017	16.0	13.6	-2.4	-14.7
2018	16.0	13.6	-2.3	-14.5

Note: The effective federal marginal tax rate on income from capital is the share of the last dollar of such income taken by federal individual income and corporate income taxes.

Taking into account the effects discussed above, CBO estimates that between 2009 and 2013, policies in the President’s proposed budget could result in either a smaller or a larger capital stock in private hands than would be predicted on the basis of CBO’s baseline. The estimated effect on the capital stock depends to a great extent on how much a change in the after-tax rate of return on saving is assumed to alter consumer spending. Under the assumption that the effect is small, the higher deficits under the President’s proposals would lead to a lower capital stock. Under the assumption that higher rates of return on saving tend to significantly reduce consumer spending, the resulting boost in saving could outweigh the effect of higher deficits, resulting in increases in the capital stock.

Some policies proposed in the President’s budget would not only affect the amount of the capital stock but would change the mix of different types of capital within that stock—a shift that also could affect potential output. The proposal in the President’s budget to extend the lower tax rates on corporate dividends and capital gains would probably have the largest effect on the composition of the capital stock because it would reduce taxes on personal income received from the corporate sector, thus encouraging a shift of some capital from the noncorporate to the corporate sector.

Currently, some corporate income is taxed once at the business level (through the corporate income tax) and again at the personal level (through the individual income tax on dividends and capital gains). (Not all corporate income is taxed in this fashion—some is effectively not taxed at the business level, and some is not taxed at the personal level.) That tax treatment affects the way that capital is allocated in the economy because it discourages investment in the corporate sector relative to investment in the housing and noncorporate business sectors. As a result, less capital may be held in the corporate sector than is optimal for the economy’s efficient operation. Extending lower tax rates on dividends and capital gains would reduce that effect, enhancing economic output.

Extending the current tax rates on dividends and capital gains also would diminish the impact of the baseline’s current-law assumption that tax rates will rise after 2010. The taxation of dividends and capital gains may encourage greater corporate reliance on debt to finance investment than would otherwise be the case. Because businesses may deduct interest payments on debt they owe (such as bonds they have issued) from any taxable income, they can lower their tax payments by borrowing instead of issuing stock to finance their investments. Interest payments thus are taxed once, at the personal level. If the current tax rates expire as assumed in the baseline, the tax treatment of interest payments on debt may influence businesses’ decisions about financing and lead to an inefficient allocation of resources.

Technological Progress. New and improved technical processes and products are the source of most long-term growth in productivity, and the President’s budgetary proposals could affect the economy by influencing the rate at which technological progress is made. Researchers, however, understand little about how taxation and spending policies affect technological innovation. Therefore, for the most part CBO has not incorporated into its analysis effects on technological progress arising from the President’s proposals.⁸

The demand-side effects of the President’s proposals would vary over time. In 2008 and 2009, increased discretionary spending would provide a boost to demand. Starting in 2010, however, the proposals would reduce discretionary spending, which tends to reduce demand, but also cut taxes, which in turn tends to encourage private consumption and thereby to enhance demand. The net effect on demand is positive in the early years and negative in the later years (CBO analyzes demand-side effects only through 2014).

In general, increases in demand may cause businesses to temporarily gear up production and hire more workers; decreases in demand may have opposite effects. From a demand-side perspective, budgetary policies that raised private and public consumption might offset some of the current slowdown in economic output. Nevertheless, demand-side effects are relatively fleeting: They can only temporarily raise or lower output beyond what it would otherwise be because stabilizing economic forces tend to move output back toward its potential level. Moreover, policies that increase demand above its potential level by raising government consumption or spurring private consumer spending are likely to lower national income in the long run because such policies eventually can tend to reduce the size of the nation’s capital stock and thus reduce national income.

The President’s budget included one proposal in particular, the "economic growth package" of tax reductions, which would tend to produce positive demand-side effects. However, that proposal is not included in this analysis because a largely equivalent measure, the Economic Stimulus Act of 2008, has already been signed into law. It is therefore included in CBO’s baseline rather than in the President’s budget.

CBO used five economic models to estimate the effects of the President’s budgetary proposals relative to the policy assumptions underlying CBO’s baseline. The models, which fall into two broad categories, focus on somewhat different aspects of the economy and reflect distinct ways of thinking about it. Three of the models estimate supply-side effects only; the other two are commercial macroeconometric models that emphasize the business-cycle aspects of the economy and are designed primarily to analyze demand-side effects, although they incorporate some supply-side influences as well. Each type of model represents individuals’ economic decisions—in particular, the degree to which individuals anticipate future developments—in an idealized way that does not capture all aspects of actual behavior. Even so, the results provide a reasonable range of estimates of individuals’ responses to changes in policy. (Figure 2-2 presents, year by year, the estimated effect of the President’s proposals on some key inputs for CBO’s various models—effective tax rates on labor and capital and the size of the deficit or surplus.)

Note: Effects on the deficit or surplus are by fiscal year; impacts on effective tax rates are by calendar year.

CBO used three growth models to analyze the supply-side effects of the President’s proposals from 2008 through 2018.⁹ The models—a "textbook" growth model, a life-cycle growth model, and an infinite-horizon growth model—differ mainly in their assumptions about how far into the future people look in making plans (see Appendix C). The textbook growth model assumes, in effect, that people do not explicitly consider expected future policies when they make plans—that is, the model incorporates no forward-looking behavior. Moreover, the model does not account for the way the changes in marginal tax rates on capital income might influence investment.

In contrast, the life-cycle model incorporates the assumption that people make lifelong plans for working and saving but do not consider events that might occur after they die. The infinite-horizon model differs yet again, assuming, in effect, that people behave as if they will live forever—or, more realistically, that they care about the well-being of their descendants as well as their own.^¹⁰ Moreover, the life-cycle and infinite-horizon models assume that people know with certainty how the government will resolve its long-term budget imbalance, whether by higher tax rates, by lower spending and transfer payments, or by some combination of the two. Both the life-cycle and the infinite-horizon models assume that households face uncertainty about future wages and could become credit constrained (that is, unable to borrow to maintain their spending) if their wages decline significantly.

CBO used the textbook growth model to estimate effects under two separate assumptions about how much people will adjust their work hours in response to changes in marginal tax rates: a "low" assumption, under which workers respond very little, and a "high" assumption, under which their response is on the high side of the consensus range of empirical estimates from studies based on one-year changes in labor supply.¹¹ CBO found that under the low assumption, the President’s proposals would decrease gross national product (GNP) by 0.2 percent, on average, over the 2009–2013 period.¹² Under the high assumption, the proposals would have little effect on GNP during those years. Over the 2014–2018 period, the proposals would reduce GNP by 0.1 percent to 0.2 percent (see Table 2-4). The effects estimated by the textbook growth model become more negative over time as effects of the proposed extension of the provisions in EGTRRA and JGTRRA on marginal tax rates on labor income are increasingly offset by the effects of the health proposal. (The model projects negative effects on the capital stock over the entire period, on average.)

		2009 to 2013	2014 to 2018


Growth Models
Without Forward-Looking Behavior
Textbook Model
	High (Hours worked respond strongly to tax-rate changes)	*	-0.1
	Low (Hours worked respond weakly to tax-rate changes)	-0.2	-0.2

With Forward-Looking Behavior
Closed-Economy Life-Cycle Model
	Government spending adjusted after 2018	0.4	0.6
	Taxes adjusted after 2018	0.4	0.4
Open-Economy Life-Cycle Model
	Government spending adjusted after 2018	0.8	1.2
	Taxes adjusted after 2018	0.7	1.0
Infinite-Horizon Model
	Government spending adjusted after 2018	0.2	0.3
	Taxes adjusted after 2018	0.2	0.2

Macroeconometric Models
Supply-Side Contribution
Macroeconomic Advisers’ Model		-0.2	n.a.
Global Insight’s Model		*	n.a.

Supply-Side and Demand-Side Contributions
Macroeconomic Advisers’ Model		-0.2	n.a.
Global Insight’s Model		-0.4	n.a.

Notes: The "textbook" growth model is an enhanced version of a model developed by Robert Solow. The life-cycle growth model, developed by CBO, is an overlapping-generations general-equilibrium model. The infinite-horizon growth model is an enhanced version of a model first developed by Frank Ramsey. The models by Macroeconomic Advisers and Global Insight, which are available commercially, are designed to forecast short-term economic developments. The various models reflect a wide range of assumptions about the extent to which people are forward-looking in their behavior: In the textbook model and those by Macroeconomic Advisers and Global Insight, people have the least foresight, whereas in the infinite-horizon model, people’s foresight is perfect and extends indefinitely to include a full consideration of effects on descendants.

For several reasons, the results of the life-cycle and infinite-horizon growth models differ from those of the textbook growth model. Unlike the textbook growth model, they are built on the assumption that individuals adjust their decisions about work, spending, and saving both in response to changes in marginal tax rates and after-tax rates of return and in anticipation of future changes in policy.

The forward-looking characteristics of the life-cycle and infinite-horizon growth models require CBO to make assumptions about what people believe will happen in the future, both in current law and under the President’s proposed policies, not only during the 10-year projection period but into the indefinite future as well. For its analysis, CBO assumed that people believe that the budgetary policies being assessed—those of the President or of CBO’s baseline—will be maintained over the entire 10‑year projection period. (In reality, people may well believe that the policies might change at some point during that time.)

For the years after 2018, however, matters are complicated by the fact that the policies reflected both in CBO’s baseline and in the President’s proposals are unsustainable in the long run, owing to projected increases in spending for health and retirement programs.¹³ To address that difficulty, CBO assumed that people expect the fiscal imbalances projected under current law to be resolved over the long run. It then made explicit assumptions about the manner in which changes in deficits or surpluses under the President’s budgetary policies, relative to those in CBO’s baseline, would eventually be reflected in spending and taxes.

The life-cycle and infinite-horizon models each generated two sets of estimates that were based on different assumptions about that financing. Under one assumption, people believe that the proposals will be financed by gradually adjusting government spending for goods and services and for transfer payments (as shares of GNP) over the 2019–2028 period. Under the other assumption, people believe that the proposals will be financed by gradually adjusting marginal tax rates over the same period.

Under either financing assumption, the infinite-horizon model projects that the President’s proposals will increase GNP by 0.2 percent over the 2009–2013 period. Depending on the assumption, the model projects increases in GNP of 0.2 percent or 0.3 percent over the 2014–2018 period. The infinite-horizon model estimates more positive effects on output than does the textbook growth model in part because it projects a larger boost in private saving when the extension of the 2001 and 2003 tax provisions increases the after-tax return on saving, leading to a higher level of the capital stock, on average. (In the textbook growth model, those higher returns are not assumed to affect private saving.) The effects of the proposals on labor supply are also somewhat more positive under the assumptions of the infinite-horizon model than under those of the textbook growth model.

The life-cycle model projects considerably more positive effects on output. Depending on which assumption about financing is used and whether the economy is considered to be open or closed to flows of foreign capital, it projects that the President’s proposals will increase GNP by 0.4 percent to 0.8 percent over the 2009–2013 period and by 0.4 percent to 1.2 percent over the 2014–2018 period. The life-cycle model estimates a larger effect on output than does the infinite-horizon model in part because its assumptions imply a greater response of labor supply to marginal tax rates, and, in most years, marginal tax rates on labor income are lower under the President’s proposals. In addition, the life-cycle model assumes that private saving and labor supply increase by a greater amount in response to reduced transfer payments (such as the reductions in Medicare spending proposed by the President) because people have a shorter horizon over which to make up for those reductions than they do in the infinite-horizon model.

The effects of the President’s proposals are smaller under the assumption of a closed economy because changes in wages and interest rates offset some of the effects of the proposals on labor supply and saving. For example, in 2011 and 2012, reduced marginal tax rates on labor encourage more hours of work, but under the closed-economy assumption the resulting addition to labor supply leads to a decline in wages, damping the effect. Similarly, in the later years of the projection, when reduced taxes on capital lead to increases in saving, growth in the capital stock causes interest rates to decline, offsetting some of the effect of taxes on the rate of return to saving. Under the open-economy assumption, by contrast, wages and interest rates are assumed always to be equal to worldwide levels and are not affected by domestic policies. The United States’ economy probably lies somewhere between the open- and closed-economy assumptions used in the life-cycle model. It is open to capital flows, but it also is large enough to influence world rates of interest and wages.¹⁴

The supply-side effects of the President’s proposed policy changes would feed back to the budget (see Tables 2-1 and 2-5). Using the assumptions underlying its baseline, CBO projects that the President’s proposals will expand the cumulative deficit over the 2009–2013 period by $336 billion, ignoring economic feedback, but that feedback could add as much as $13 billion to the total or subtract as much as $151 billion from it, depending on which set of assumptions is used in the analysis. For the period from 2014 to 2018, the President’s budgetary policies are projected to boost the cumulative deficit by $651 billion, ignoring economic feedback, but that feedback could add as much as $21 billion to the increase or subtract as much as $193 billion from it. No single number is likely to provide an accurate measure of the feedback, but the numbers presented here illustrate the range of its probable magnitude.

		2009 to 2013	2014 to 2018


Growth Models
Without Forward-Looking Behavior
Textbook Model
	High (Hours worked respond strongly to tax-rate changes)	8	-4
	Low (Hours worked respond weakly to tax-rate changes)	-13	-21

With Forward-Looking Behavior
Closed-Economy Life-Cycle Model
	Government spending adjusted after 2018	70	150
	Taxes adjusted after 2018	64	116
Open-Economy Life-Cycle Model
	Government spending adjusted after 2018	151	193
	Taxes adjusted after 2018	135	155
Infinite-Horizon Model
	Government spending adjusted after 2018	35	72
	Taxes adjusted after 2018	35	68

Macroeconometric Models
Supply-Side Contribution
Macroeconomic Advisers’ Model		-29	n.a.
Global Insight’s Model		*	n.a.

Supply-Side and Demand-Side Contributions
Macroeconomic Advisers’ Model		-76	n.a.
Global Insight’s Model		-38	n.a.

Notes: Numbers in this table reflect the effects on the cumulative deficit or surplus of the economic effects shown in Table 2-4. (Negative numbers indicate an increase in the deficit; positive numbers, a reduction.) They do not include the estimated costs of the President’s budgetary proposals under CBO’s baseline economic assumptions. The total effect of the proposals on the cumulative deficit or surplus, including both those direct costs and the secondary effects shown above, appear in Table 2-1.

Because demand-side economic developments become increasingly hard to estimate the farther projections extend into the future, CBO analyzed demand-side effects of the President’s budgetary proposals only for the first five years of the 2009–2018 period. To do so, it used macroeconometric forecasting models created by two private forecasting concerns—Macroeconomic Advisers and Global Insight. Each model includes an embedded growth model, but each concentrates primarily on demand-side economic effects.

As with the textbook growth model, CBO adjusted the models of Macroeconomic Advisers and Global Insight to incorporate its own estimates of how workers would adjust their hours worked in response to the changes in marginal tax rates on labor income implied by the President’s proposals.

Like the textbook growth model, Macroeconomic Advisers’ and Global Insight’s models are not forward-looking—people, as the models represent them, do not behave as though they have specific expectations about future policies or economic developments. However, the models do represent individuals as responding to some economic changes in the same way that they have responded in the past, regardless of the source of those changes. For example, people are assumed to react to proposals to change marginal income tax rates and after-tax labor income in roughly the same way they did, on average, when after-tax wages and income changed in the past.

The lack of forward-looking behavior in the macroeconometric models implies that specific policy changes that are scheduled to occur in the future will not affect current behavior unless special adjustments are made to mimic such behavior.¹⁵ For example, the President’s proposals would reduce taxes throughout the projection period. Those lower taxes would increase the amount of after-tax income that people expected in the future, which might cause them to boost their spending today (as the forward-looking models imply). In the macroeconometric models, however, those changes in taxes affect consumer spending only when they occur.

CBO explored the relative magnitude of demand- and supply-side effects of the proposed policies by adjusting monetary policy responses in the models. For one set of scenarios, CBO assumed that the Federal Reserve would respond to economic developments in a standard way that would accommodate demand- and supply-side effects.^¹⁶ For a second set, CBO assumed that the Federal Reserve would respond in such a way as to hold the unemployment rate as projected in CBO’s baseline. The second approach produced an estimate of the implications of the proposals for potential (noncyclical) GNP—in other words, the supply-side effects. Subtracting the second set of results from the first provides an estimate of the demand-side effects of the proposed policies.¹⁷

Incorporating CBO’s estimate of effects on labor supply, the Macroeconomic Advisers’ model predicted that the demand- and supply-side effects of the President’s proposed policies would decrease GNP by 0.2 percent, on average, between 2009 and 2013 (see Table 2-4). Global Insight’s model forecast a decrease of 0.4 percent. The Macroeconomic Advisers’ model projected that the supply-side effects of the President’s proposals would decrease output by 0.2 percent over the 2009–2013 period, whereas the Global Insight model projected, on average, little supply-side effect on output.

Both models conclude that the proposals’ projected economic impacts would feed back to the budget and affect the size of the projected deficit. According to the projections from the Macroeconomic Advisers’ model, feedback effects on the supply side could add $29 billion to the $336 billion increase in the deficit projected for the 2009–2013 period under the CBO baseline’s economic assumptions (see Table 2-5). By the estimates of Global Insight’s model, the supply-side feedback of the President’s proposals over the same period would have little effect on the deficit.

On the demand side, the proposals’ feedback effects on the budget would be similar in the two models. In the Macroeconomic Advisers’ model, the proposals’ effect on demand would increase the deficit by about $47 billion over the 2009–2013 period, yielding an overall feedback from demand- and supply-side effects of raising the deficit by $76 billion over the period. In Global Insight’s model, negative demand-side effects lead to an overall feedback effect that raises the deficit by $38 billion from 2009 to 2013. The overall increase in the deficit is greater as estimated by the Macroeconomic Advisers’ model than it is by Global Insight’s model, despite a smaller estimated decline in output, because the Macroeconomic Advisers’ model projects an increase in interest rates, on average, which raises federal interest payments, whereas Global Insight’s model projects a decrease in interest rates.

Increased deficits and the attendant increases in interest payments must be offset by policy changes at some point, or interest costs would compound relative to output over time, driving the debt-to-output ratio ever higher (under the assumption, which CBO’s findings incorporate, that the rate of interest on government debt is higher than the rate of economic growth).

See Congressional Budget Office, Labor Supply and Taxes (January 1996). Since that report was published, CBO has revised downward its estimates of total wage elasticity and substitution elasticity for secondary earners because of evidence that their responsiveness has declined over time as their participation in the labor force has grown. Also see Francine D. Blau and Lawrence M. Kahn, "Changes in the Labor Supply Behavior of Married Women: 1980–2000," Journal of Labor Economics, vol. 25, no. 3 (2007), pp. 393–438.

CBO’s current estimate of the effect on marginal tax rates of the proposal to change the tax treatment of health insurance premiums is initially smaller than last year’s because this year’s estimate assumes that the effect on marginal rates would phase in slowly. The proposal affects marginal tax rates only to the extent that people are able to adjust the quantity of employer-provided health insurance they receive. This year’s estimate assumes that people are able to make that adjustment only slowly, whereas last year’s estimate assumed they could make it right away.

Both sets of estimates yield effective tax rates that are below all but the lowest statutory marginal rates because some capital income (for example, interest income that flows into tax-free savings accounts or pension funds and rental income from owner-occupied housing) is not taxed. The expansion of tax-free savings accounts was analyzed differently. See Appendix B for details.

CBO used two commercial macroeconometric models to estimate the demand-related effects of the President’s proposals. Global Insight’s model has potential GDP responding positively to spending for research and development—which in turn would be stimulated by the proposal to extend tax credits for such activities.

Growth models are often called "supply-side models." They assume that the labor market is always in equilibrium (that fiscal policy has no effect on the unemployment rate). CBO presents effects for the 2009–2013 and 2014–2018 periods because the main purpose of this discussion is to illustrate how economic feedback could affect the budget numbers presented in Chapter 1 for those periods. The models showed positive supply-side effects on output in 2008 that stemmed mainly from reduced tax rates on labor income as a result of the proposal to extend for one year the increase in the exemptions allowed under the alternative minimum tax. The positive effects in 2008 are reflected in calculations of revenues, government debt, and interest costs for 2009 and beyond.

This year CBO incorporated a new assumption about income variability in the infinite-horizon model. On the basis of academic research, CBO increased the year-to-year variation in income that people are assumed to face (see David Domeij and Jonathan Heathcote, "On the Distributional Effects of Reducing Capital Taxes," International Economic Review, vol. 45, no. 2 (May 2004), pp. 531–532). That change tends to reduce the estimated effect of policy actions (such as changes in marginal tax rates).

CBO’s estimates used data from a large sample of taxpayers to account for the effects of changes in marginal tax rates and in after-tax income under the President’s proposals. The models incorporated a larger response to changes in marginal tax rates among secondary earners than among primary earners.

In presenting the economic effects of the President’s budgetary proposals, CBO uses gross national product as its measure of output rather than the more commonly cited gross domestic product. Changes in GNP exclude foreigners’ earnings on investments in the domestic economy but include domestic residents’ earnings overseas and are therefore a better measure of the proposals’ effects on domestic residents’ income than are changes in GDP in an open economy like that of the United States. The budget calculations presented in Table 2-5 reflect the fact that tax treaties and other factors result in some foreign income being effectively untaxed.

The infinite-horizon model assumes a closed economy. The textbook growth model and the models of Macroeconomic Advisers and Global Insight make assumptions that are effectively intermediate between the life-cycle model’s open- and closed-economy assumptions.