[Page: H3829]


   The SPEAKER pro tempore. Under a previous order of the House, the gentleman from California (Mr. THOMAS) is recognized for 5 minutes.

   Mr. THOMAS. Mr. Speaker, pursuant to clause 3 (h)(2)(A)(iii) of rule XIII, I submitted the following macroeconomic impact analysis:

   In accordance with House Rule XIII.3(h)(2), this document, prepared by the staff of the Joint Committee on Taxation (``Joint Committee staff''), provides a macroeconomic analysis of H.R. 2, the ``Jobs and Growth Reconciliation Tax Act of 2003.'' The analysis presents the results of simulating the changes contained in H.R. 2 under three economic models of the economy. The models employ a variety of assumptions regarding Federal fiscal policy, monetary policy, and behavioral responses to the proposed changes in law.

   1. Description of Models and Results Format


   The Macroeconomic Equilibrium Growth (``MEG'') model.--This model, developed by the Joint Committee staff, is based on the standard, neoclassical assumption that the amount of output is determined by the availability of labor and capital, and in the long run, prices adjust so that demand equals supply. This feature of MEG is comparable to a Solow growth model, described as the ``textbook growth model'' by the Congressional Budget Office (An Analysis of the President's Budgetary Proposals for Fiscal Year 2004, March 2003, pp. 28-29) (``CBO''). Individuals are assumed to make decisions based on observed characteristics of the economy, including current period wages, prices, interest rates, tax rates, and government spending levels. Because individuals do not anticipate changes in the economy or government finances, this type of behavior is referred to as

[Page: H3830]
``myopic behavior.'' Consumption in MEG is determined according to the life-cycle theory, which implies that individuals attempt to even out their consumption patterns during their lifetimes.

   MEG differs from a simple neoclassical growth model in that prices in MEG adjust to equilibrate supply and demand with a delay or lag, rather than instantaneously. This feature allows the model to simulate a disequilibrium adjustment path, in which resources may be underemployed or over-employed (used at an unsustainable rate) in response to policies that stimulate or depress economic activity. It also allows an analysis of the effects of differing intervention policies by the Federal Reserve Board. In this respect, the MEG model resembles econometric models such as the Macroeconomic Advisers model and the Global Insight model.

   In the MEG simulations in each of the tables below, it is assumed that the Federal Reserve Board either acts aggressively by raising interest rates to counteract almost completely any demand stimulus provided by H.R. 2 (``MEG aggressive Fed response''), or remains neutral with respect to any changes in fiscal policy, allowing temporary changes in demand to affect levels of employment and output (``MEG neutral Fed response'').

   The Global Insight (``GI'') econometric model.--Like the MEG model, this commercially available model is capable of simulating disequilibrium adjustments to changes in demand. The model is made up of a set of equations that estimate from historical data the behavioral coefficients that determine the timing and strength of economic relationships within the model. Comparable parameters in the MEG and OLG models are derived from economic research. In many cases this research is also based on econometric analysis of historical data.

   Individuals and firms behave myopically in the GI model. For this analysis, the Joint Committee staff uses an estimated monetary reaction function designed to moderate gradually, but not completely offset, deviations from full employment by lowering or increasing interest rates. Thus, if the economy is operating near capacity, proposals that increase employment and accelerate the economy will result in increasing interest rates.

   The overlapping generations life cycle model (``OLG'').--In this model, individuals are assumed to make consumption and labor supply decisions with perfect foresight of economic conditions, such as wages, prices, interest rates, tax rates, and government spending, over their lifetimes. The OLG model is similar to the type of model described as a ``life cycle model'' by the CBO, ibid.

   One result of the perfect foresight assumption is that if a policy results in an economically unstable outcome, such as increasing government deficits indefinitely into the future, the model will not solve. Therefore, to run simulations in this model, it is necessary to assume that an offsetting budget balancing fiscal policy will be enacted. In the tables below, it is assumed that either government spending will be reduced after 2013 to offset the tax cut (``OLG future government spending offset'') or individual income tax rates will be increased after 2013 (``OLG future tax rate increase'').

   The cut in government spending to offset the costs of a tax cut can be modeled either as a cut in transfer payments, as is presented here, or as a cut in ``non-productive government spending.'' The latter assumption is used in CBO, ibid. The difference between the two approaches is that consumers are assumed to value transfer payments, and thus work and save more within the budget window in anticipation of losing them; but they are assumed not to value non-productive spending, and therefore do not increase work or savings in anticipation of this cut. Thus, the anticipation of valued spending cuts results in more growth in the early years than the anticipation of non-valued spending cuts.


   Because the exact time path of the economy's adjustment to changes such as a new tax policy is highly uncertain, the Joint Committee staff presents results as percent changes during the Congressional budgeting time frame. In addition, for the MEG and OLG models, which have been designed to provide long-run equilibrium results, information is provided about the long run. While it is impossible to incorporate unknowable intervening circumstances, such as major resource or technological discoveries or shortages, these models are designed to predict the long-run effects of policy changes, assuming other, unpredictable influences are held constant.

   Because the MED model is myopic, if the policy simulated is ultimately a fiscally unstable policy, such as a net decrease in taxes that produces deficits that grow faster than the rate of growth of the economy, ``long-run'' is defined as the last period before the model fails to solve because of this unstable situation. For the OLG simulations, which incorporate a stabilizing fiscal policy offset, ``long-run'' is defined as the eventual steady-state solution.

   2. Estimated Macroeconomic Effects of H.R. 2

   The magnitude of the macroeconomic effects generated by these simulations depends upon a number of assumptions, some of which are described above, that are inherent in the models used. Several additional assumptions detailed below.


   Effect of tax rate reductions on investment.--Reductions in marginal tax rates (tax rates on the last dollar of income earned) on interest, dividend, or capital gains income create incentives for individuals to save and invest a larger share of their income, as each additional dollar of investment yields more after-tax income. Conversely, reductions in the average tax rate on income from capital provide taxpayers with more after-tax income for the same amount of investment, reducing their incentive to save and invest. Changes in the statutory tax rate affect both marginal and average rates of tax on these sources of income, providing potentially offsetting incentives. Consistent with existing research, the model simulations assume that on net, the marginal rate effect is slightly larger than the average rate effect, and thus decreases in tax rates on capital income increase savings.

   Effect of reductions in the dividend tax rate.--There is general agreement that dividend taxation reduces the return on investments financed with new share issues. However, there are two alternative views regarding the effect of dividend taxation on corporate investment returns financed with retained earnings. The ``traditional view'' holds that reductions in dividend taxes would lower the cost of corporate investment financed with either new share issues or retained earnings, and thus would provide an incentive for corporations to increase investment. Alternatively, the ``new view,'' holds that a reduction in the dividend tax rate would not lower the cost of corporate investment financed with retained earnings. Under this view, a decrease in the dividend tax rate would result in an immediate increase in the value of outstanding stock reflecting the reduction in dividend tax payments, thus increasing the wealth of the stockholders, and providing an incentive for additional consumption. The model simulations assume that half of the corporate sector is in accordance with the traditional view and half with the new view.

   Foreign investment flows.--Increased Federal government budget deficits increase the amount of borrowing by the Federal government. Unless individuals increase their savings enough to finance completely the increased deficit, the increase in government borrowing will reduce the amount of domestic capital available to finance private investment. This effect is often referred to as the ``crowding out'' of private business activity by Federal government activity. A reduction in national saving may lead to a reduction in domestic investment, and domestic capital formation, depending on the mobility of international capital flows. The government and private firms would compete for the supply of available funds and interest rates would rise to equate the demand and supply of funds. Returns on foreign investments would accure mainly to foreigners and would only increase the resources available to Americans to the extent that higher domestic investment resulted in higher wages in the United States. The MEG and GI simulations incorporate an assumption that there would be some in-flow of foreign capital to the extent that the rate of return on capital is increased by the tax policy. However, the inflow in foreign capital is not enough to offset completely the increased Federal borrowing. The OLG simulations assume there is no inflow of foreign capital.

   Effect of tax rate reductions on labor supply.--As in the case of savings responses, tax rate reductions provide offsetting labor supply incentives. Reductions in the marginal tax rate on earnings create an incentive to work more because taxpayers get to keep more of each dollar earned, making each additional hour of work more valuable; while reductions in the average tax rate create an incentive to work less, because they result in taxpayers having more after-tax income at their disposal for a given amount of work.

   Consistent with existing research, the simulations assume that taxpayers in different financial positions respond differently to these incentives. Typically, the largest response comes from secondary workers (individuals whose wages make a smaller contribution to household income than the primary earner in the household) and other underemployed individuals entering the labor market. As described above, labor supply responses are modeled separately for four different groups in MEG: low income primary earners, other primary earners, low income secondary earners, and other secondary earners.

   Effects of reductions in tax liability on demand.--Generally, any net reduction in taxes results in taxpayers making more purchases because they have more take-home income at their disposal. Policies that increase incentives for taxpayers to spend their income rather than save it provide a bigger market for the output of businesses. The amount of economic stimulus resulting from demand side incentives depends on whether the economy has excess capacity at the time of enactment of the policy, and on how the Federal Reserve Board reacts to the policy. If the economy is already producing near capacity, demand-side policies may, instead, result in inflation, as consumers bid up prices to compete for a fixed amount of output. If the Federal Reserve Board believes there is a risk that the policy will result in inflation, it may raise interest rates to discourage consumption. In this case, depending on how strongly the Federal Reserve Board

[Page: H3831]
reacts, little, if any increase in spending will occur as a result of would-be stimulative tax policy. The MEG aggressive Fed response simulation assumes the Federal Reserve Board completely counteracts demand stimulus; the MEG neutral Fed response simulation assumes the Federal Reserve Board ignores the stimulus; and the GI simulation assumes the Federal Reserve Board partially counteracts demand stimulus. The OLG simulations have no monetary sector because they assume demand automatically adjusts to supply through market forces.


   Economic Growth.--

  Calendar years 
  2003-08   2009-13 
Neoclassical Growth Model:  
MEG--aggressive Fed reaction   0.3   0.2  
MEG--neutral Fed reaction   0.9   1.0  
Econometric Model:  
GI Fed Taylor reaction function   1.5   1.2  
Life Cycle Model With Forward Looking Behavior:  
OLG Reduced Government Spending in 2014   n.a.   n.a.  
OLG Increased Taxes in 2014   n.a.   n.a.

  Calendar years 
  2003-08   2009-13 
Neoclassical Growth Model:  
MEG--aggressive Fed reaction   0.2   -0.1  
MEG--neutral Fed reaction   0.3   0.0  
Econometric Model:  
GI Fed Taylor reaction function   0.9   -0.1  
Life Cycle Model With Forward Looking Behavior:  
OLG Reduced Government Spending in 2014   0.2   -0.1  
OLG Increased Taxes in 2014   0.2   -0.2

   As shown in Table 1, depending on the assumed Federal Reserve Board reaction to the policy, the estimated change in Gross Domestic Product (``GDP'') due to this proposal can range at least from a 0.3 percent (an average of $43 billion) to a 1.5 percent (an average of $183 billion) increase in nominal, or current dollar GDP over the first five years, and 0.2 percent to a 1.2 percent increase over the second five years. As shown on Table 2, depending on the assumed Federal Reserve Board reaction to the policy, and on how much taxpayers anticipate and plan for the effects of future Federal government deficits, the change in real (inflation-adjusted) GDP due to those proposal can range from a 0.2 percent (an average of $18 billion per year) to a 0.9 percent (an average of $76 billion per year) increase in real GDP over the first five years, with a small decrease over the second five years.


  Calendar years 
  2003-08  2009-13 
Percent Change in Non-Residential Capital Stock


Neoclassical Growth Model:  
MEG--aggressive Fed reaction   0.6   0.4  
MEG--neutral Fed reaction   0.8   0.6  
Econometric Model:  
GI Fed Taylor reaction function   1.5   0.4  
Life Cycle Model With Forward Looking Behavior:  
OLG Reduced Government Spending in 2014   0.1   -0.7  
OLG Increased Taxes in 2014   0.1   -0.8


Percent Change in Residential Housing Stock


Neoclassical Growth Model:  
MEG--aggressive Fed reaction   -1.0   -1.5  
MEG--neutral Fed reaction   -0.8   -1.1  
Econometric Model:  
GI Fed Taylor reaction function   -0.5   -1.3  
Life Cycle Model With Forward Looking Behavior:  
OLG Reduced Government Spending in 2014   -0.2   -0.1  
OLG Increased Taxes in 2014   -0.2   -0.1

   As the results in Table 3 indicate, this policy may increase investment in non-residential capital in the first five years by 0.1 percent to 1.5 percent, while reducing investment in residential capital by -0.2 percent to -1.0 percent because of the reduced cost of capital, which is due to the reduction in taxation of dividends and capital gains, and the temporary bonus depreciation. The investment incentives for producers' equipment in this proposal are likely to shift some investment from housing to other capital. The size of the shift differs between the simulations because of different assumptions about adjustment costs and savings responses. In the second five years, the sunset of the bonus depreciation provision, combined with the negative effects of crowding out will slow increases in private nonresidential investment. The simulations indicate that eventually the effects of the increasing deficit will outweigh the positive effects of the tax policy, and the build up of private nonresidential capital stock will likely decline.

   Labor Supply and Employment.--

  Calendar years 
  2003-08  2009-12 
Neoclassical Growth Model:  
MEG--aggressive Fed reaction   0.2   0.0  
MED--neutral Fed reaction   0.4   -0.1  
Econometric Model:  
GI Fed Taylor reaction function   0.8   -0.4  
Life Cycle Model With Forward Looking Behavior:  
OLG Reduced Government Spending in 2014   0.2   -0.1  
OLG Increased Taxes in 2014   0.2   -0.1

   As shown in Table 4, employment may increase from 0.2 percent (approximately 230,000 new jobs) to 0.8 percent (about 900,000 new jobs) in the first five years, as the effects of the acceleration of individual rate cuts, and the initial increase in investment prevail. Employment increases in the first five years because of both the positive labor supply incentive from the individual rate cuts, and the economic stimulus effect of the proposal taken as a whole. This increase disappears by the end of the budget period, ranging from 0 percent to -0.4 percent. The acceleration of the individual tax rate reductions is effectively a temporary provision relative to present law; thus, the positive labor supply incentives are temporary.

   A substantial portion of the tax cuts in the proposed growth package, those attributable to the acceleration of the individual income tax provisions in the Economic Growth and Tax Relief Reconciliation Act of 2001 (``EGTRRA''), and the bonus depreciation/NOL carryback combination are temporary (operating from 2003-2006), and therefore likely to result in modest demand stimulus primarily in the first five years in the myopic models. In the OLG stimulations, in which individuals foresee the temporary nature of the stimulus, the increase in consumption is spread across both periods.

   3. Budgetary Effects

   When the macroeconomic effects of a change in tax policy are taken into account, estimates of the change in receipts due to the proposal may change. To the extent that a new policy changes the rate of growth of the economy, it is likely to change the amount of taxable income, which will have a ``feedback effect'' on receipts. In addition, by increasing the after-tax return on investments in capital that generate taxable income, a change in policy may shift investment from non-taxable or tax-favored sectors, such as the owner-occupied housing market, into the taxable sector, and thereby increase receipts. The model simulations indicate that the policy analyzed here is likely to result in more economic growth in the first five years than under current law, and hence results in less revenue loss than what is predicted using conventional revenue estimates. As the GDP growth declines in years 6-10, the revenue feedback also declines.

   A change in policy, however, may result in inflation as well as real economic growth. Inflation causes increases in nominal revenues (revenues measured in current dollars), without necessarily increasing the purchasing power of the Federal government. Conventional budget analysis is conducted in nominal dollars. To the extent that this analysis applies equally to revenue and expenditure estimates, this practice provides a reasonably accurate picture of the effects of inflation on the Federal budget. However, the Joint Committee staff analyzes the effects of tax policy on receipts, but not spending. Reporting revenues due to inflation, without reporting the commensurate budget effects would present an inaccurate picture of the effects of the proposal on the entire deficit. Therefore, the Joint Committee staff provides budgetary analysis in real (inflation-adjusted), rather than nominal terms. Table 5 shows the percent revenue feedback relative to the conventional revenue estimate, in real terms.

   Even when presented in real terms, revenue feedback analysis alone may provide an incomplete picture of the effects of tax policy on the Federal budget. To the extent that the policy results in a net decrease in Federal receipts, with no offsetting expenditure reductions, the policy results in an increase in the Federal deficit. Increases in the Federal deficit generate additional debt service costs.

   To determine how changes in tax policy affect the ability of the government to meet its current and future obligations it is helpful to compare tax-induced changes in the deficit and GDP. If GDP is growing faster than the deficit, the fiscal situation is improving, whereas if the deficit is growing faster, the fiscal situation is worsening. If deficits are growing faster (slower) than GDP, then the ratio of Federal debt to GDP would increase (decrease), which implies that future generations would have less (more) income to consume and invest after making payments on the debt.

  Calendar Years  
  2003-08   2003-13 
Neoclassical Growth Model:  
MEG--aggressive Fed reaction   9.8   3.6  
MEG--neutral Fed reaction   27.5   23.4  
Economic Model:  
GI Fed Taylor reaction function   16.1   11.8  
Life Cycle Model With Forward Looking Behavior:  
OLG Reduced Government Spending in 2014   6.1   3.0  
OLG Increased Taxes in 2014   5.8   2.6

   Table 5 shows the relationship between the change in receipts generated using macroeconomic analysis, and the predicted change in receipts provided by a conventional revenue estimate. A positive percentage indicates the estimated revenue loss is less when macroeconomic effects are taken into account than when estimated using conventional methods. As the simulations indicate, depending on how much temporary demand stimulus is generated by the proposal, the revenue feedback could range from 5.8 percent to 27.5 percent in the first five years, and 2.6 percent to 23.4 percent over the ten-year budget period.

   4. Data Sources

   All of the macroeconomic models used by the Joint Committee staff are based primarily on quarterly National Income and

[Page: H3832]
Product Account (``NIPA'') data published by the Bureau of Economic Analysis, U.S. Department of Commerce. In the MEG model, and to the extent possible in the commercial models, Joint Committee staff use the forecast for Federal and State and local government expenditures and receipts forecast by the Congressional Budget Office (The Budget and Economic Outlook: Fiscal Years 2004-2013, January 2003) instead of the NIPA series for these fiscal variables. For purposes of modeling changes in average and marginal tax rates in the macroeconomic models, the Joint Committee staff use microsimulation models that are based on tax return data provided by the Statistics of Income Division of the Internal Revenue Service (``SOI'').

   The Joint Committee staff uses these microsimulation models to determine average tax rates and average marginal tax rates for the different sources of income in each model, and to calculate the changes in these rates due to the proposal. The tax calculator calculates the change in liability due to the proposal for each record. These changes are aggregated for use in the macroeconomic models according to the different levels of disaggregation in each model. In the aggregations, averages are weighted by the income for each group. The percent change in average and marginal rates due to this proposal are:

Year  Average tax rate on wages  Average marginal tax rate on 
    Wages  Interest  Dividends  Capital gains 
2003   -11   -9   -11   -51   -24  
2004   -10   -6   -8   -49   -23  
2005   -9   -3   -6   -52   -24  
2006   0   0   0   -48   -23  
2007   -1   0   0   -48   -23  
2008   0   0   0   -50   -22  
2009   -1   0   0   -47   -22  
2010   -1   0   0   -48   -22  
2011   -1   0   0   -52   -22  
2012   -1   0   0   -50   -21  
2013   0   0   0   0   0

   To obtain information about the effects of proposals affecting business tax liability, the Joint Committee staff uses a corporate tax microsimulation model that is similar in structure to the individual tax model. This data source for the corporate model is a sample of approximately 140,000 corporate tax returns provided by SOI.

   Depending on the requirements of the policy simulation, the corporate model can be run either on a full cross section of sampled tax returns, (i.e., one full year, or on a panel of returns constructed from any combination of tax years in the 1987 through 1998 period). This panel feature is particularly useful in tracking net operating losses and credits that can be either carried back or carried forward to other tax years.

   Finally, Joint Committee microsimulation tax calculators are also used to help assess the effect of a tax proposal on the cost of capital because some firms are operating at or near a net operating loss (``NOL'') position, not all of the 50 percent of equipment expenses can be deducted by each firm each year. A key component of the cost of capital is the net present value of depreciation deductions. An increase in the value of the depreciation deduction lowers the cost of capital. The calculated percent increases in the net present value of the depreciation deduction due to this proposal are shown below (the change is different for each of the first three years because of the temporary nature of the bonus depreciation provisions in present law and in the proposal):

Year   Percent change from present law 
2003   8.3  
2004   9.1  
2005   15.4  
2006   .005

   5. Conclusion

   The Joint Committee staff model simulations indicate that H.R. 2 would likely stimulate the economy immediately after enactment by creating temporary incentives to increase work effort, business investment, and consumption. This stimulus is reduced over time because the consumption, labor, and investment incentives are temporary, and because the positive business investment incentives arising from the tax policy are eventually likely to be outweighted by the reduction in national savings due to increasing Federal government deficits.