Key Points
Legislation like the 2022 CHIPS Act provides sector-specific tax incentives. However, these types of sector-specific investment incentives have received very little attention in the literature using models with realistic fiscal effects and where prices are determined in general equilibrium.
Using a multi-sector version of the PWBM dynamic OLG model, we examine the economic and budgetary effects of two major reforms targeting C-corporations in the computer and electronic product manufacturing sector: a 10 percent production credit and a 10-percentage-point increase in business expense deductions.
As expected, both proposals increase investment and output in the target computer sector. However, both policies have very small effects on the overall U.S. economy level of output, even though the targeted sector is a production complement that effectively increases the marginal total factor productivity of other sectors.
Computer Sector Investment Tax Incentives: Economic and Budgetary Effects
Legislation like the 2022 CHIPS Act provide sector-specific incentives, in this case, to the production of U.S. semiconductors. However, there is very little understanding of how sector-specific investments actually boost economywide production, such as GDP, within a comprehensive overlapping-generations model with realistic fiscal externalities and prices computed in general equilibrium.
This paper provides a high-level description of the multi-sector version of PWBM’s Dynamic OLG model. For illustration, we show how these model improvements allow us to evaluate two tax policies directed at the computer and electronic product manufacturing sector: a 10 percent production credit and a 10-percentage-point increase in business expense deductions.
PWBM has developed a multi-sector model of production to analyze sector-specific tax proposals. Although international tax changes can affect the performance of U.S. business sectors, this brief focuses on how changes to domestic business tax policies designed to expand the production in the computer and electronic product manufacturing sector (“Computer Sector”) affects that sector, other sectors in the U.S. economy, the U.S. economy as a whole, and the federal debt.
In the Dynamic OLG Model, firms are defined by their production sector—the type of good that they produce—and their business type—whether they operate as C-corporations or pass-through entities. Each firm has its sector- and type-specific tax schedule and production parameters. Changes to these firm-specific tax schedules influence a firm’s decisions to produce or invest, which in turn affects interest rates, wages, and other prices. Those price changes affect household decision-making as well as government taxation and spending in PWBM’s Dynamic OLG Model. By rigorously modeling sector-specific business decisions, PWBM can comprehensively assess the response of households, the aggregate economy, and the federal budget to sector-specific tax changes.
Although PWBM’s Dynamic OLG model is a very flexible tool for economic analysis, the model may not be able to incorporate all potential legislative objectives. For example, the CHIPS Act emphasizes the importance of maintaining a supply of semiconductors to support U.S. national security interests, a non-pecuniary externality that is not yet measured within PWBM’s framework. We leave that consideration up to future work.
PWBM models domestic production by dividing it into individual sectors, each represented by firms operating in competitive product markets. These firms use labor and capital as inputs, for which they pay market-rate wages and rents. The output from each sector is an intermediate good, and all intermediate goods are combined into a final good using a Cobb-Douglas production function. That final good can be used for household consumption, government consumption, or capital investment. For example, consider two sectors: automobile manufacturing and gasoline production. In this example, cars and gas combine to create a final good: travel. Changes in prices, productivity, or taxes on either of these two goods can affect the total number of trips produced and the amount of tax revenue.
The Dynamic OLG model is designed to use however many sectors are needed to address a specific policy question. Domestic production can be modeled as a single sector for the entire U.S. economy or as numerous sectors of different intermediate goods. Each sector is characterized by different tax policies and three key production parameters. Each sector has its own tax policy, which includes sector-specific tax, credit, deduction, and depreciation schedules.
The three key production parameters are income share (the amount of GDP that comes from that sector), labor share (the amount of income that goes to wages in each sector), and economic depreciation (the rate at which productive capital wears out and becomes unproductive). Sectors like healthcare are far more labor-intensive than sectors such as semiconductor manufacturing. More labor-intensive sectors are characterized by a higher labor share, meaning that a larger portion of income is allocated to wages rather than to rents on capital assets such as factories, buildings, or equipment. Economic depreciation also varies across sectors—buildings lose their productive value much more slowly than assets like trucks or computer equipment.
PWBM’s Dynamic OLG model distinguishes between two types of businesses within each sector: C-corporations and pass-through entities (S-corporations, partnerships, and sole proprietorships). These business types are taxed differently. C-corporations face double taxation, with income taxed first at the corporate level and again at the owner level when profits are distributed. In contrast, pass-through income is taxed only at the owner level.
In the Dynamic OLG model, labor and capital are distributed across sectors so that investors earn the same returns—net of business taxes—on each unit of capital invested in C-corporations, regardless of sector. This means that, after all taxes are accounted for, a hypothetical investor building a production facility or purchasing equipment in a C-corporation will receive the same return on their investment, no matter which sector they invest in. Since the returns on capital in C-corporations must be equal across sectors, any change to tax, credit, and deduction schedules in one sector causes investors to reallocate their investments, which influences capital, investment, labor, and production in all sectors, not just the one in which the changes occur. A fixed proportion of each sector’s labor and capital is allocated between C-corporations and pass-through businesses.
In each period (model years), businesses pay wages to households for providing labor and pay returns to those who invest their savings in the firm, which is used as productive capital. C-corporations choose their levels of labor, capital, and debt based on prevailing market rates to maximize profits, which are distributed to households that invest in them. These corporations are subject to a flat 21 percent tax on their taxable income. However, their effective tax rate is often lower due to various deductions and credits.
Many firms choose to issue debt, as interest payments are mostly tax deductible. In the Dynamic OLG model, the cost of issuing debt increases as the amount of debt grows relative to the firm's size. Larger firms, with more capital available as collateral, face lower borrowing costs than smaller firms because their capital reduces lender risk. Smaller firms encounter higher costs for issuing the same amount of debt. Firms can significantly reduce their taxable income through deductions for interest payments on debt.
Firms can further lower their tax liability through various deductions from gross income. The largest deduction typically covers employee compensation, including wages and benefits. C-corporations can take a range of other available deductions that usually account for 10 to 15 percent of gross income. Pass-through businesses typically deduct close to 18 percent.1 In one of the illustrative policy experiments, PWBM increases the value of that deduction by 10 percentage points for C-corporations, simulating aimed at encouraging greater investment in the workforce and allowing additional expense deductions.
The Dynamic OLG model also incorporates a comprehensive treatment of capital depreciation allowances by sector. Firms that make capital investments can deduct a fraction of that investment from taxable income over 40 years. Adjusting the depreciation schedule can affect the present value of current investments. Allowing larger deductions earlier in the schedule increases the net present value of the tax deduction and makes investments more attractive.
Finally, PWBM incorporates a sector-specific production credit into its model. That credit provides firms with a tax reduction based on a fraction of the value of their output, encouraging increased production. In our second illustrative policy experiment, we show how increasing the production credit can incentivize firms in specific sectors to increase their output.
For our illustrative analysis, we divide the economy into two production sectors: one for the Computer Sector and another representing the rest of the economy.2 We choose parameters for each sector that allow the model to reproduce important economic features of each sector.
In 2022, the Computer Sector generated approximately $360 billion worth of output, accounting for about 1.8 percent of non-farm business GDP. The parameters for the Cobb-Douglas production function for the final good output are set so that the Computer Sector is about 1.8 percent of the economy in the Dynamic OLG.
Data from the Bureau of Economic Analysis (BEA) shows that the labor share for the Computer Sector is 57 percent and the labor share for the rest of the economy excluding the Computer Sector is 62 percent. In this case, the Computer Sector uses a slightly greater share of capital equipment in its production process than the rest of the U.S. economy does.
The labor shares in the Dynamic OLG are set to the same values for each of these two sectors. According to the BEA, firms in the Computer Sector pay about 1.8 percent of all wages and the sector uses about 1.9 percent of the productive capital.
We set the capital depreciation rate for the Computer Sector to 7.5 percent per year, about 2 percentage points higher than that for the rest of the economy. With these assumptions, the Dynamic OLG model aligns the capital usage in the Computer Sector with the figure reported in BEA data.
The share of income related to pass-through businesses in the Computer Sector is much lower than it is for the rest of the economy: Approximately 80 percent of income goes to C-corporations in the former, compared to about 61 percent in the rest of the economy.
Under current law, the statutory rate for C-corporations is 21 percent and 0 percent for pass-through businesses. Those rates are the same in our two production sectors.
PWBM estimates that credits that are related to output are somewhere around 0.8 percent for C-corporations in the broader economy and typically vary between 1.4 to 1.5 percent for C-corporations in the Computer Sector, depending on the year. As pass-through companies do not pay business taxes, their credits here are set to zero. Interest rate deductibility is typically a little above 85 percent for all firms regardless of business type or sector.
Other deductions from taxable income start at about 10 percent in 2023 and decline to about 6 percent by 2060 for C-corporations in the computer and electronic product sector.
For pass-through businesses in the rest of the economy, the deduction rate is a bit higher. These deductions start just under 15 percent in 2023 and slowly decline to about 9 percent in 2060.
We consider two experiments to evaluate how changes to the tax code in one sector—in this case, the Computer Sector—affect production in that sector as well as the whole of the U.S. economy.
The first experiment is a permanent 10-percentage-point increase in the deductions for C-corporations in the Computer Sector. This could reflect proposals such as those that would allow corporations additional deductions for expenses such as additional workforce training and education.
The second experiment provides C-corporations with a 10 percent credit for production in the Computer Sector. For every dollar worth of the intermediate good produced, the government refunds 10 cents back to the firm. Like the other illustrative policy, this credit is also permanent.
A 10-percentage-point increase in the deductions for C-corporations in the Computer Sector increases the returns to business capital within that sector, which encourages additional investment. The result is an increase in capital available in that sector. As the top panel in Table 1 shows, over the 2030 to 2045 period, capital increases by 2.5 percent relative to current law. That increase means an additional $43 billion in productive capital is used in the Computer Sector in 2030. Together with a small 0.1 percent decrease in total effective labor used in this sector, total output increases by 1 percent compared to current law.
| 2030 | 2035 | 2040 | 2045 | |
|---|---|---|---|---|
| Computer and Electronic Product Manufacturing Sector | ||||
| Capital | 2.5% | 2.5% | 2.5% | 2.5% |
| Labor | -0.1% | -0.1% | -0.1% | -0.1% |
| Output | 1.0% | 1.0% | 1.0% | 1.0% |
| Total Wages | 0.0% | 0.0% | 0.0% | -0.1% |
| All Other Non-Farm Business Sectors | ||||
| Capital | -0.1% | -0.1% | -0.1% | -0.1% |
| Labor | 0.0% | 0.0% | 0.0% | 0.0% |
| Output | 0.0% | 0.0% | 0.0% | 0.0% |
| Total Wages | 0.0% | 0.0% | 0.0% | 0.0% |
The bottom panel of Table 1 shows how this policy change affects the rest of the economy. Because the increase in deductions increased the return to capital in the Computer Sector, investments are reduced in the rest of the economy and redirected to the Computer Sector until the expected returns between the two sectors are equalized. However, because the rest of the economy makes up over 98 percent of the economy, the small shift in capital to the Computer Sector decreased in capital in the rest of the economy by only 0.1 percent. As a result, changes to other aggregates, such as labor supply and output, are very small in the rest of the economy.
| 2030 | 2035 | 2040 | 2045 | |
|---|---|---|---|---|
| Gross domestic product | 0.0% | 0.0% | 0.0% | 0.0% |
| Capital stock | 0.0% | 0.0% | 0.0% | 0.0% |
| Hours worked | 0.0% | 0.0% | 0.0% | 0.0% |
| Average wage | 0.0% | 0.0% | 0.0% | 0.0% |
| Government debt | 0.1% | 0.2% | 0.2% | 0.2% |
The overall effects on the economy, shown in Table 2, are similarly small. Overall, government debt increases by between 0.1 percent in 2030 and by 0.2 percent in 2045. The larger deduction leads to a decline in federal revenues from business taxes, which leads to higher government debt. On one hand, higher government borrowing crowds out investment in productive capital, which leads to a smaller stock of economy-wide productive capital. On the other hand, the increase in deductions increases the average return by lowering business taxes on productive capital. These effects mostly offset each other, which leads to an economy that otherwise looks similar to the baseline economy.
Although the policy has helped the overall economy to shift production into the Computer Sector, total output, capital, hours worked, and the average wage do not change in a meaningful way.
| 2030 | 2035 | 2040 | 2045 | |
|---|---|---|---|---|
| Computer and Electronic Product Manufacturing Sector | ||||
| Capital | 11.8% | 11.9% | 11.9% | 11.9% |
| Labor | 11.9% | 11.9% | 11.9% | 11.9% |
| Output | 11.9% | 11.9% | 11.9% | 11.9% |
| Total Wages | 12.1% | 12.2% | 12.1% | 12.1% |
| All Other Non-Farm Business Sectors | ||||
| Capital | -0.3% | -0.2% | -0.3% | -0.3% |
| Labor | -0.2% | -0.2% | -0.2% | -0.2% |
| Output | -0.2% | -0.2% | -0.2% | -0.2% |
| Total Wages | 0.0% | 0.0% | 0.0% | 0.0% |
A 10-percentage-point production credit for C-corporations in the Computer Sector offers a 10-cent credit for every dollar of production generated. As the top panel of Table 3 shows, this policy leads to a large influx of investment into the Computer Sector, which leads to an increase in productive capital of just under 12 percent over the 2030-45 period, equivalent to an increase in the capital stock of about $226 billion in 2030. Labor and output move in lockstep with capital: total effective labor and the production of computers and electronics increases by slightly less than 12 percent. Total wages paid by the sector increase by slightly more than 12 percent in all four reported years.
The bottom panel of Table 3 illustrates the impact that policy change has on the rest of the economy. As profitability improves in the Computer Sector, investments are redirected towards that sector, but the decrease in the rest of the economy is small due to its much larger size. Capital decreases by between 0.2 and 0.3 percent, labor supply and output decline by 0.2 percent, and the change in the total wage bill is negligible.
| 2030 | 2035 | 2040 | 2045 | |
|---|---|---|---|---|
| Gross domestic product | 0.0% | 0.0% | 0.0% | 0.0% |
| Capital stock | 0.0% | 0.0% | 0.0% | 0.0% |
| Hours worked | 0.0% | 0.0% | 0.0% | 0.0% |
| Average wage | 0.2% | 0.2% | 0.2% | 0.2% |
| Government debt | 0.8% | 1.2% | 1.4% | 1.6% |
Table 4 shows that the overall effect of this policy change on the U.S. economy is small. The cost of the policy, $43 billion per year in 2030, is much larger than the 10-percentage point deduction and leads to an increase in government debt of about 0.8 percent in 2030 and 1.6 percent in 2045. As under the previous policy, the increase in government debt crowds out investment in productive private capital, which offsets the incentive to invest provided by the higher returns from the production credit. The overall effect on total hours worked, the capital stock, and GDP is close to zero. Nonetheless, this policy has a small positive effect on the wage. The wage rate increases by 0.2 percent in each reported year.
This analysis was produced by Jon Huntley under the direction of Felix Reichling and the faculty director, Kent Smetters. Mariko Paulson prepared the brief for the website.
-
Deductible expenses include rent and lease payments, utilities, interest on business loans, depreciation, supplies and materials, advertising and marketing, legal and professional fees, travel and meals, certain taxes, insurance premiums, and training and education. ↩
-
Specifically, we parameterize the computers and electronic products sector based on the BEA's Computer and Electronic Product Manufacturing subsector, classified under NAICS 334. ↩