Financing corporate tax cuts with shareholder taxes

Anagnostopoulos, Alexis; Atesagaoglu, Orhan Erem; Cárceles-Poveda, Eva Article Financing corporate tax cuts with shareholder taxes Quantitative Economics Provided in Cooperation with: The Econometric Society Suggested Citation: Anagnostopoulos, Alexis; Atesagaoglu, Orhan Erem; Cárceles-Poveda, Eva (2022) : Financing corporate tax cuts with shareholder taxes, Quantitative Economics, ISSN 1759-7331, The Econometric Society, New Haven, CT, Vol. 13, Iss. 1, pp. 315-354, https://doi.org/10.3982/QE1167 This Version is available at: https://hdl.handle.net/10419/296276 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc/4.0/ Quantitative Economics 13 (2022), 315–354 1759-7331/20220315 Financing corporate tax cuts with shareholder taxes Alexis Anagnostopoulos Department of Economics, Stony Brook University Orhan Erem Atesagaoglu Faculty of Business, Istanbul Bilgi University Eva Cárceles-Poveda Department of Economics, Stony Brook University We study the aggregate and distributional consequences of replacing corporate profit taxes with shareholder taxes, namely taxes on dividends and capital gains, in a setting with incomplete markets and heterogeneity at both the household and the firm level. The reform yields distributional gains with a large majority of households benefiting. Moreover, if dividend and capital gains are taxed at the same rate, the reform is also efficiency-enhancing and the implied optimal corporate income tax rate is zero. In contrast, an asymmetric tax treatment of dividend and capital gains induces a trade-off between efficiency and distributional concerns that is optimally resolved at a positive optimal corporate tax rate, implying double taxation. Keywords. Optimal corporate taxes, double taxation, heterogeneity, misallocation. JEL classification.E6. 1. Introduction Corporate income tax cuts remain one of the most polarizing topics in fiscal policy. This issue returned to the forefront of political debate recently with the Tax Cuts and Jobs Act of 2017, which included a sizable reduction in corporate profit rates. Often, proponents of the tax cuts emphasize the inefficiency of raising revenues using corporate taxes relative to other income taxes, while opponents argue that the revenue loss induced by the reforms would have to be compensated with personal income tax hikes or cutbacks in benefits programs, targeted at the least wealthy, if the reforms are to be revenue-neutral. Alexis Anagnostopoulos: [email protected] Orhan Erem Atesagaoglu: [email protected] Eva Cárceles-Poveda: [email protected] This paper was previously circulated under the title “On the Double Taxation of Corporate Profits.” We wish to thank Arpad Abraham, Juan Carlos Conesa, Allan Drazen, Ayse Imrohoroglu, Ayse Kabukcuoglu, Andrea Lanteri, Han Ozsoylev, Joseph Zeira as well as seminar participants at the EUI, Koc, Southampton, St. Andrews, Sabanci, Bilgi and Bogazici Universities and conference participants at the Midwest Macro, SED, CRETE, and North American Summer meetings for helpful comments and suggestions. We also thank two anonymous referees for suggestions that have improved the paper substantially. ©2022 The Authors. Licensed under the Creative Commons Attribution-NonCommercial License 4.0. Available at http://qeconomics.org.https://doi.org/10.3982/QE1167 316 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) The academic literature provides ample support to both the positive efficiency gains from lower corporate taxes and the potential negative distributional effects.1 In this paper, we study whether a corporate profit tax reform can deliver some of the efficiency gains from a corporate tax cut while, at the same time, avoiding the negative distributional consequences and gaining popular support. We address these questions in an infinite horizon framework with incomplete markets that features idiosyncratic uncertainty at both the household and the firm level. In such a setting, we show that corporate profit tax cuts can gain widespread political support whenever revenueneutrality is induced via higher taxes that fall on the same group of people, namely, the shareholders. To be more specific, we consider dividend and capital gains taxes and investigate whether increasing one, or both, of them to compensate for a reduction in the corporate tax can lead to efficiency, distributional, and overall welfare improvements. In addition, by considering a series of tax cuts of different size, we are able to determine the optimal mix of corporate and shareholder taxes. This also sheds light on the question of whether double taxation of dividends is justified from an optimal perspective. To our knowledge, our model is the first one to investigate these issues in a setting with a substantial amount of both household and firm heterogeneity. From a pure efficiency perspective, our analysis can be thought of as a comparison between the relative importance of the distortions caused by the corporate tax versus the distortions caused by shareholder taxes. First, we argue that the answer can be misleadingly simple in the context of a standard growth model with no heterogeneity. In that context, corporate income taxes reduce investment incentives by lowering the after tax returns to investment, capital gains taxes also distort investment by raising the cost of capital, but a constant dividend tax does not distort the investment decision because it does not directly affect the returns to investment (although it affects stock prices).2 This would suggest that concentrating all taxes on dividends only would be the optimal choice. However, this conclusion is unwarranted when markets are incomplete. When households face uninsurable idiosyncratic risk, the wealth effect arising from the stock price changes is transmitted in general equilibrium to savings and investment, implying that the neutrality of dividend taxes is no longer true. In addition, when firms seek external financing to grow, a difference between the dividend tax rate and the capital gains tax rate acts as a financing friction and leads to distortions in the allocation of capital across firms.3One of the main objectives of this paper is to quantify and compare the direct distortions of the corporate profit tax with the indirect distortions of shareholder taxes in the presence of a tax wedge between the dividend and capital gains tax rates. 1See, for example, the literature based on the classic Chamley–Judd results and more recent work in incomplete markets setups such as Domeij and Heathcote (2004) and Conesa, Kitao, and Krueger (2009). 2See McGrattan and Prescott (2005), Santoro and Wei (2011), and Atesagaoglu (2012) among many others. 3These two points are made in Anagnostopoulos, Carceles-Poveda, and Lin (2012), in the context of a model with only household heterogeneity, and by Gourio and Miao (2010), in the context of a model with only firm heterogeneity. The two papers study the effects of a reduction in shareholder taxes, but they do not study changes in the corporate tax rate. Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 317 The preceding discussion suggests that the distortions due to the tax wedge could potentially be avoided simply by increasing the capital gains tax in tandem with the dividend tax, which avoids introducing the wedge. However, this would introduce a direct distortion of the capital gains tax on the cost of capital and it is an open question whether this distortion compares favorably to the direct one caused by corporate taxes. We argue that, in a simple growth model, these distortions are identical and the corporate tax is equivalent to an equal tax on dividends and capital gains. Moreover, we provide conditions under which this result can be extended to an economy with incomplete markets and external financing, a result that constitutes a theoretical contribution in itself. However, the equivalence between corporate and shareholder taxes relies on a definition of taxable corporate income, which is at odds with the actual tax code. In our quantitative analysis, which uses a more standard definition of taxable income, the equivalence is no longer true. Clarifying and quantifying the direct distortions of corporate and shareholder taxes is another important objective of the present paper. In addition to efficiency considerations, we are interested in the distributional effects of the reforms so our model incorporates household heterogeneity. As a result, our model features both a continuum of households that are subject to uninsurable idiosyncratic labor income risk and a continuum of firms that are subject to idiosyncratic productivity shocks. Firms use decreasing returns to scale technology that combines labor and capital to produce output. They own capital directly and decide on investment, payout, and financing policy. The latter consists in choosing between using internal funds or issuing new equity. Households can trade in shares of a mutual fund comprising all firms and earn asset income, in the form of dividends and capital gains from their share holdings, as well as labor income. The government maintains a fixed amount of exogenous spending, which it can finance through flat taxes on firms’ profits and on households’ labor and asset income. Starting at the benchmark calibrated economy, we consider permanent changes in the corporate tax rate and concurrent increases in shareholder taxes that maintain long run government revenue fixed. In the first experiment, we increase both dividend and capital gains taxes maintaining the equality between the two. In the second experiment, only dividend taxes are increased and this introduces a tax wedge between dividend and capital gains taxes.4In both experiments, wages increase and capital returns decrease in the long run. This ensures that households at the bottom of the wealth distribution, that rely mainly on labor income, benefit from the reforms. Thus both types of reform have positive distributional consequences, in the sense that high marginal utility households benefit, and are supported by a large majority of households. Interestingly, this stands in contrast to corporate tax cuts financed through labor taxes, which tend to imply negative redistribution and limited support. At the same time, the two reforms are markedly different regarding their effects on efficiency. When only dividend taxes are increased, we show that the resulting misallocation of capital due to the wedge in shareholder taxes dominates the distortions caused by the 4When dividend taxes are increased above capital gains taxes, firms have an incentive to use repurchases instead of dividends to avoid the dividend tax and this is not allowed in our model. As a result, the policy change considered here implicitly includes the introduction of a rule against repurchases. 318 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) corporate tax. Although aggregate capital and output increase significantly due to the corporate tax reduction, the misallocation of capital combined with large transitional costs due to the short run increase in savings and drop in consumption lead to welfare losses from an aggregate perspective. Using a utilitarian social welfare function, these aggregate losses are traded off against the positive distributional effects. For large reductions in the corporate tax rate, social welfare decreases because the aggregate component dominates, while smaller reductions have a quantitatively small, positive effect on social welfare. The implication is that social welfare is maximized at a positive corporate tax rate, implying that double taxation can be an optimal response to the efficiency versus distribution trade-off in this case. In contrast, increasing both dividend and capital gains taxes together yields both efficiency and distributional benefits. These become larger, the larger the decrease in corporate taxes, which means that the optimal choice would be to eliminate corporate taxes in this case. The efficiency benefits arise due to an improvement in capital allocation. In the long run, aggregate capital is lower but more efficiently distributed, leading to higher output. In contrast to the standard effects of capital tax cuts, which induce additional savings to increase long run output, the transition here features a reduction in savings and an increase in consumption, which generates positive efficiency effects. Overall, our results suggest that a reform, which maintains equality of dividend and capital gains taxes might be preferable in the sense that it delivers efficiency gains on top of the distributional gains. Although eliminating corporate taxes while maintaining the equality of shareholder taxes yields the highest social welfare gains, it would represent a dramatic change that might be politically difficult to implement in practice. A less dramatic corporate tax reduction would be to consider a reform which equalizes the tax rates for all types of personal income as well as for corporate income. We include results from such an experiment, where the common tax rate required is approximately 28%, and we find that such a reform would lead to overall welfare gains and command wide political support in the sense of welfare gains for 84% of households. The reform which maintains equality of dividend and capital gains taxes is also more robust to relaxing the assumption that tax changes are unexpected. We show this by also computing transitions and welfare under the assumption that the reform is anticipated 1 or 2 years in advance. In that case, a reform that increases only dividend taxes can have very different implications regarding the short run responses of macroeconomic aggregates because firms engage in tax arbitrage in an attempt to take advantage of the temporarily low dividend tax. This tax arbitrage has the effect of introducing additional fluctuation in wages during the transition and this mostly affects low-wealth individuals. As a result, the distributional benefits of the reform are reduced. Given the computational complexity involved,5the model necessarily abstracts from several other potentially important mechanisms through which corporate taxes can affect macroeconomic outcomes. Recent studies have identified some of those 5The double-sided heterogeneity is further complicated by the presence of occasionally binding constraints for both firms and households as well as the need to go further than steady states and compute transition paths in order to evaluate the welfare consequences of reforms. Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 319 mechanisms, such as the importance of the choice of the legal form of organization (Chen, Qi, and Schlagenhauf (2018)), the presence of lumpy investment (Miao and Wang (2014)) or the role of capital mobility in an open economy setting (Fehr, Jokisch, Kambhampati, and Kotlikoff (2013)). None of these studies consider shareholder taxation as part of the suggested reform and this is where our paper’s contribution lies relative to them. Motivated by the Jobs and Growth Tax Relief Reconciliation Act of 2003, Gourio and Miao (2010)andAnagnostopoulos, Carceles-Poveda, and Lin (2012) investigate the effects of reducing shareholder taxes, but are silent about changes in the corporate profits tax. Relative to the former, our model incorporates household heterogeneity and incomplete markets, which are crucial in order to capture the effects of shareholder taxes on precautionary savings as well as to evaluate the distributional welfare effects of tax reforms. Relative to the latter, our model incorporates firm heterogeneity and external financing, which are crucial in order to evaluate the distortionary effects of an increase in dividend taxes. Integrating both mechanisms within the same framework is important since they can have opposite implications regarding the effects of shareholder taxes. Conesa and Dominguez (2013) is most closely related to our work, since it investigates corporate taxes in conjunction with dividend taxes. They show that the optimal scheme in the long run features zero corporate taxes and positive dividend and labor income taxes that are equalized to each other. While they go one step further by computing optimal Ramsey taxes rather than once and for all tax rate changes, they abstract from capital gains taxes and heterogeneity, implying that their model does not capture the distortions arising from the tax wedge in shareholder taxes when markets are incomplete. Incorporating those elements, we show that switching from corporate taxes to dividend taxes is only a welfare improving policy if capital gains taxes are also increased. Section 2provides the model, Section 3discusses the main qualitative insights, Section 4presents the calibration of the benchmark economy, Section 5presents the quantitative results, and Section 6discusses the sensitivity of our results to modeling and calibration choices. Section 7concludes. 2. The model We consider an infinite horizon economy, where time is discrete and indexed by t.Idiosyncratic firm productivity shocks generate firm heterogeneity and, at the same time, idiosyncratic labor efficiency shocks generate household heterogeneity. Both types of shocks wash out in the aggregate so that there is no aggregate uncertainty in this model. To keep the model tractable, we assume households trade only a single asset, which is interpreted as a mutual fund composed of all the firms in the economy as in Favilukis, Ludvigson, and van Nieuwerburgh (2017). A government maintains a balanced budget every period by taxing firm profits as well as household labor, dividend, and capital gains income. 320 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) 2.1 Households There is a continuum (measure 1) of households indexed by iwith identical utility functions given by E0 ∞  t=0 βtu(cit ), where β∈(0, 1)is the subjective discount factor, cit denotes consumption, and E0de- notes the expectation conditional on information at date t=0. The period utility function u(·):R+→Ris assumed to be strictly increasing, strictly concave and continuously differentiable, with limci→0u(ci)=∞and limci→∞ u(ci)=0. In the absence of leisure in the utility, households supply a fixed amount of labor (normalized to one) and receive labor income that is exogenous from their point of view. The economy wide real wage rate is denoted by wtbut each household is subject to an idiosyncratic shock it to their productivity, so that labor income of household iis wtit . The productivity shock is i.i.d. across households and follows a Markov process with transition matrix (|)and Npossible values. Markets are incomplete. Households can only partially insure against uncertainty by trading shares θit of a mutual fund, which comprises all the firms in the economy. Holding shares provides income to the household in the form of dividends as well as capital gains resulting from changes in the market value of these shares. Since there is no aggregate uncertainty, dividends and share prices are certain and the traded asset is risk-free. Households face proportional taxes on labor income, dividend income, and capital gains income at rates of τlt,τd,andτg, respectively. They can use their after tax income from all sources to purchase consumption goods or to buy shares θit of the mutual fund at a competitive market price Pt. After tax income includes labor income and the income from holding shares θit−1. These shares entitle the household to a share θit−1of the total after tax dividend payout (1−τd)Dt. In addition, the shareholder can sell their shares at apriceP0 t, which represents the time tvalue of equity outstanding in period t−1. The increase in the value of this existing equity (P0 t−Pt−1)θit−1represents accrued capital gains, which are taxed at the rate τg.6Since we allow firms to raise new equity St,the market value of equity at time t(after new equity is issued) is Pt=P0 t+St. The households’ budget constraint can be expressed as cit +Ptθit =(1−τlt )wtit +(1−τd)Dt+P0 tθit−1−τgP0 t−Pt−1θit−1.(1) Short selling of the mutual fund shares is not allowed: θit ≥0. 6We make the simplifying assumption that capital gains taxes are paid on an accrual basis and that capital losses are subsidized at the same rate. This is the standard approach in the literature with the notable exceptions of Gavin, Kydland, and Pakko (2007) and Dammon, Spatt, and Zhang (2001,2004). Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 321 In each period t, households choose how much to consume and how many shares to buy given prices, dividends and tax rates {Pt,P0 t,wt,Dt,τlt,τd,τg}∞ t=0. The optimal consumption/savings choice is described by a standard Euler equation, which holds with equality for unconstrained households, 1+rt+1≡1+(1−τd)Dt+1+(1−τg)P0 t+1−Pt Pt =u(cit ) βEtu(cit+1),(2) where we have defined the net after tax return to be rt+1. Note that, given the absence of aggregate uncertainty, that return is deterministic. Equation (2) simply states that, at an optimum, the after tax return on the asset must equal the intertemporal marginal rate of substitution of unconstrained households. It can be used to express the market value of the mutual fund as a present discounted sum of tax adjusted payouts Pt= ∞  j=1j  k=1 1 1+rt+k 1−τg (1−τd) (1−τg)Dt+j−St+j.(3) Note that shareholders discount future payouts using the before tax return r 1−τgand that equity issuance reduces the payout for current shareholders. When τd=τg,thepayout is simply Dt+j−St+jbut when τd>τ g, then dividends are valued less than capital gains. 2.2 Firms The production sector follows Gourio and Miao (2010) with some modifications. Firms use capital kand labor lto produce consumption goods yusing a Cobb– Douglas production function with decreasing returns to scale y=zf (k,l)=zkαklαl, where 0 <α k,αl<1andαk+αl<1. Production is subject to an idiosyncratic productivity shock zwhichisi.i.d.acrossfirmsandfollowsaMarkovprocesswithtransition matrix z(z|z)and Nzpossible values. We now consider the problem of a particular firm j. Each period t, given the available capital and the current productivity realization, firm jchooses labor demand optimally. The choice of labor demand is a static problem and it defines the operating profit of the firm as follows: π(kjt,zjt;wt)≡max ljt zjtf(kjt,ljt )−wtljt, where wtis the economy wide wage rate. The firm’s labor demand is determined by the following optimality condition: wt=αlzjtkαk jt lαl−1 jt . Given the determination of operating profits, we can now turn to the dynamic aspect of the firm’s decision making problem, which includes the investment, financing, 322 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) and payout decisions. The firm has two sources of funds, internal and external. External funds are obtained by issuing new equity.7The value of new equity issued in period t is denoted by sjt. Internal funds consist of operating profits π(kjt,zjt;wt)net of taxes τcTjt,whereTjt denotes taxable income and τcis a flat corporate income tax rate τc. Funds can be allocated to dividends djt or capital expenditures, the latter consisting of new additions to the capital stock xjt and capital adjustment costs (xjt,kjt ).Thus,the firm’s financing constraint is given by djt +xjt +(xjt,kjt )=π(kjt,zjt;wt)−τcTjt +sjt, where Tjt =π(kjt,zjt;wt)−δkjt −φ(xjt,kjt ).(4) Deductions from taxable income include a depreciation allowance δkjt as well as a fraction φof adjustment costs. The firm’s capital stock evolves according to kj,t+1=xjt +(1−δ)kjt,(5) where δ∈[0, 1]is the capital depreciation rate. Finally, we assume dividend payments cannot be negative djt ≥0(6) and no repurchases are allowed8 sjt ≥0. (7) We assume that firm jmaximizes the following objective: E0 ∞  t=0t  n=1 1 1+rn 1−τg 1−τd 1−τg djt −sjt based on the mutual fund value in (3). Specifically, the discount factor used by all firms is the risk-free rate. Although individual firms are not directly traded here, the discount factor used is consistent with firms being traded as long as households hold diversified portfolios, that is, as long as no individual firm jis a large enough proportion of a household’s portfolio to affect its marginal rate of substitution.9 7In Section 6, we extend the model to also allow for debt financing. 8This assumption is innocuous for the calibrated versions of our model where τd=τg.Forthecases where dividend taxes are raised above capital gains taxes, we refer the reader to Gourio and Miao (2010) for a discussion of the relevance of the assumption as well as the potential effects from relaxing it. For additional discussion, see also Section 7. 9The mutual fund assumption is borrowed from Favilukis, Ludvigson, and van Nieuwerburgh (2017)who focus on the housing market, specifically the variability of the price-rent ratio. In their model, there are only two firm-sectors, a consumption good producing sector and a housing sector. Households buy stocks in a mutual fund that combines these two productive sectors. In their model, a household’s MRS can covary with sector specific (aggregate) shocks making the choice of firm objective nontrivial. We refer the reader to their paper for a discussion of alternative assumptions regarding the discount factor in that context. Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 329 Table 1. Parameter values—baseline calibration.a Parameter Value Discount Factor β0.934 Share of Capital in Production αk0.311 Share of Labor in Production αl0.650 Depreciation Rate δ0.054 Adjustment Cost Parameter ψ1.210 CRRA Parameter μ1.00 Fraction of Adjustment Cost Deducted ϕ0.52 Labor Productivity Shocks it See Table 2 Firm Level Productivity Shocks zit See Table 4 Tax Rate on Corporate Income τc0.34 Tax Rate on Dividends τd0.20 Tax Rate on Capital Gains τg0.20 Tax Rate on Labor Income τl0.28 aSee “Section 4: Calibration” for details on data resurces used in the becnhmark calibration. chain model of only three states.15 The highest productivity is almost 50 times the lowest productivity and the transition matrix is such that the stationary distribution of labor earnings has 50% of households at the low productivity, 44% with medium productivity and only 6% with high productivity. Together, these imply a Gini coefficient of labor earnings of 0.6 just as in US data. Importantly, the transition matrix is constructed so that it induces (endogenously) a highly skewed wealth distribution. The idea is that there is a substantial risk of dropping from the top of the earnings distribution to the bottom and this induces top earners to accumulate large precautionary savings. In our benchmark economy, the Gini coefficient of wealth is 0.87, which is only slightly larger than the value of 0.816 reported in Diaz-Gimenez, Glover, and Rios-Rull (2011). Quintiles of the wealth distribution are reported in Table 3, together with their counterpart in the data.16 As the table shows, more than 90% of wealth is held by the top quintile both in the data and in the model. Table 2. Household labor productivity process.a =[1.00 5.29 46.55 ] ∗ =[0.498 0.443 0.059 ] (/)=0.992 0.008 0.000 0.009 0.980 0.011 0.000 0.083 0.917  aNotation: denotes the values of the labor productivity shock, ∗ is the stationary distribution of the labor productivity shock process, and (/)is the Markov transition matrix. 15For details on this, see also Diaz, Pijoan-Mas, and Ríos-Rull (2003) and Castañeda, Diaz-Gimenez, and Ríos-Rull (2003). 16The data is taken from Abraham and Carceles-Poveda (2010) who use data for net financial assets from the 2004 Survey of Consumer Finances (SCF). Net financial assets exclude residential property, vehicles, and direct business ownership from the assets as well as mortgages and vehicle loans from the liabilities. 330 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) Table 3. The distribution of financial wealth in the model and the data.a Economy Quintiles Q1 Q2 Q3 Q4 Q5 Model 0.00 0.00 0.00 2.72 97.28 Data −1.55 0.09 1.61 8.66 91.20 aData for net financial assets from Abraham and Carceles-Poveda (2010) based on the 2004 Survey of Consumer Finances (SCF). Net financial assets exclude residential property, vehicles and direct business ownership from the assets as well as mortgages and vehicle loans from the liabilities. The depreciation rate δis set to 0.054 following Atesagaoglu (2012) who computes this using National Income and Product Accounts and Fixed Asset Tables data for the post WWII period. For the production function and firm productivity shocks, we use the calibration from Gourio and Miao (2010). They estimate the degree of decreasing returns to scale using Compustat Industrial Annual Data. The production function parameters αkand αlare obtained by choosing αl=0.650 to match the average labor income share in US data and αk=0.311 to capture the estimated degree of decreasing returns to scale. The process for firm level productivity shocks is estimated by fitting an AR(1) process to the residuals ztof their estimated regression: ln zt=ρln zt−1+εt,εt∼N0, σ2. The estimated values for ρand σare 0.767 and 0.211, respectively. This process is approximated using a 10 state Markov chain, shown in Table 4, obtained by applying the method of Tauchen and Hussey (1991). Finally, the adjustment cost function is assumed to be (x,k)=ψ 2(x k−δ)2kand the parameter ψ=1.210 is chosen to match the cross sectional volatility of investment rates of 0.156 in Compustat data from 1988 to 2002, as reported in Gourio and Miao (2010).17 Regarding government variables, we set the labor income tax rate to τl=0.28 following Mendoza, Razin, and Tesar (1994).18 For shareholder taxes, we use τd=τg=0.20, which is the top statutory rate in effect since the American Taxpayer Relief Act of 2012 for a vast majority of households.19 We follow Gourio and Miao (2010) in setting the corporate tax rate τc=0.34, which is roughly consistent with the statutory rate at the top bracket (0.35). Given those tax rates, government budget balance implies a value of G= 0.186, which means that government revenues are 28% of output Yin the stationary distribution. This data counterpart is closer to the model’s available liquid asset than the more traditional net worth definition. A similar picture emerges when net worth is used instead. 17Cooper and Haltiwanger (2006) find this volatility to be 0.337 in the Longitudinal Research Database (LRD). We use the value from Compustat since that data set focuses on publicly traded corporations, which are the relevant entities for our corporate tax experiments. We consider sensitivity of our results with respect to this choice in Section 6. 18Using the same methodology, but more recent data, Domeij and Heathcote (2004) report a similar value. 19The values of 20% are consistent with the 2013 federal average marginal income taxes on qualified dividends and long term capital gains reported by Feenberg and Coutts (1993). Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 331 Table 4. Firm level productivity process.a z=[0.36 0.47 0.59 0.73 0.90 1.11 1.36 1.69 2.13 2.79] ∗ z=[0.00 0.02 0.08 0.16 0.24 0.24 0.16 0.08 0.02 0.00] z(z/z)= ⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ 0.308 0.463 0.195 0.031 0.003 0.000 0.000 0.000 0.000 0.000 0.062 0.327 0.404 0.175 0.030 0.002 0.000 0.000 0.000 0.000 0.007 0.114 0.354 0.360 0.141 0.022 0.002 0.000 0.000 0.000 0.001 0.022 0.166 0.374 0.316 0.106 0.014 0.001 0.000 0.000 0.000 0.003 0.045 0.218 0.385 0.269 0.073 0.007 0.000 0.000 0.000 0.000 0.007 0.073 0.269 0.385 0.218 0.045 0.003 0.000 0.000 0.000 0.001 0.014 0.106 0.316 0.374 0.166 0.022 0.001 0.000 0.000 0.000 0.002 0.022 0.141 0.360 0.354 0.114 0.007 0.000 0.000 0.000 0.000 0.002 0.030 0.175 0.404 0.327 0.062 0.000 0.000 0.000 0.000 0.000 0.003 0.031 0.195 0.463 0.308 ⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ aNotation: zdenotes the values of the firm level productivity shock, ∗ zis the stationary distribution of the firm level productivity shock process, and z(z/z)is the Markov transition matrix. Auerbach (1989) argues that, even though capital costs such as installation costs are treated as capital expenditures in US tax law and are therefore not immediately deductible, they nevertheless generate deductions in the future through depreciation allowances. He shows how one can incorporate the present value of these deductions as immediate deductions and we follow that approach in Appendix Bto obtain a reasonable value for the fraction φof adjustment costs that can be immediately deducted from corporate taxes. Using a steady state approximation, we obtain a present value of depreciation allowances using the expression δ r 1−τg+δ. In the benchmark version of our model, we set φ=0.52, which is the value implied by this expression in the prereform stationary distribution.20 Note that, if some part of the adjustment costs cannot be attributed to investment,21 then these costs would be immediately deductible and the value of φ would be higher than what we have assumed. Given this, and the fact that φ=0.52 is only an approximation that is specific to our calibration, we also present results for the alternative extreme case with φ=1inTable7. Sinceweassumethatτd=τgin the benchmark economy, firms can be in one of the following two financing regimes: the dividend distribution (DD) regime or the equity issuance (EI) regime. Firms in the DD regime have sufficient internal funds to cover their desired level of investment, they do not need to issue equity and they pay the residual cash flow as dividends. These are typically firms with low marginal product, either due to low ztor due to high capital. In contrast, firms with high marginal product will typically need to issue equity to grow and will be in the EI regime. A third financing regime discussed in Gourio and Miao (2010), liquidity constraint firms (LC), is not present in the 20We only use a steady-state approximation because allowing for time variation in the fraction of deductions would introduce an additional state variable significantly complicating our numerical solution. Note also that we do not take into account the changes induced by endogenous changes in τgand rin our experiments, since this has a quantitatively small impact on our results. 21For example, this would be the case if such costs represent lower productivity due to worker retraining to handle new equipment. 332 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) Table 5. Distribution of firms across finance regimes (data vs. model) benchmark economy— (pre-reform steady state). Equity Issuance Regime Liquidity Constrained Regime Dividend Distribution Regime Share of Capital Dataa0.21 0.06 0.73 Model 0.19 0.00 0.81 Earnings/Capital Ratio Dataa0.56 0.29 0.33 Model 0.35 n/a 0.13 Tobin’s Q Dataa3.63 1.81 2.50 Model 1.96 n/a 1.20 aThe data reported are authors’ calculations using COMPUSTAT Industrial Annual data for the years 1988–2006. Firms that simultaneously issue equity and distribute dividends are classified under the “Equity issuance Regime”. Their share of capital is 17%. benchmark economy. However, these firms will exist post-reform whenever the reform introduces a tax wedge τd>τ g. In that case, equity issuance is costly and some firms with intermediate levels of marginal product will not find it optimal to pay the cost and will instead grow internally without paying dividends. Table 5provides some of the characteristics of the distribution of firms across the EI and DD regimes in the benchmark. The table displays the share of capital, the earnings to capital and the average Tobin’s Q for each of the regimes, together with their data counterpart.22 Consistent with the data, EI firms in the model are relatively small, have higher earnings to capital ratios, and higher Tobin’s Q. Most of the capital in the economy is held by firms in the DD regime and the share of capital held across the different regimesisconsistentwiththedata. Finally, Table 6shows features of the investment rate distribution in our benchmark economy and compares to the data reported in Cooper and Haltiwanger (2006). Table 6. Moments of the investment rate distribution. DataaModel Inactive (|x k|<0.01) 0.081 0.061 Positive Spike ( x k>0.2) 0.186 0.161 Negative Spike ( x k<−0.2) 0.018 0.003 Positive Investment ( x k>0.01) 0.815 0.577 Negative Investment ( x k<−0.01) 0.104 0.360 aData Source: Cooper and Haltiwanger (2006). 22We use Compustat annual data between 1988 and 2006 and we follow the standard criteria described in Gourio and Miao (2010) to clean the data and construct the variables. Whenever firms distribute dividends and issue equity at the same time, something that is not possible in our model, we classify these firms as equity issuance firms. Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 333 As is well known, the data shows evidence of lumpy investment. Specifically, a significant fraction (19%) of firms experience positive spikes in investment rates, a significant fraction (8%) are inactive and there is an asymmetry in that more firms undertake positive investment than negative investment. Although our model does not include non convex costs, the combination of the estimated idiosyncratic shock process together with the calibrated convex adjustment costs can take us a long way toward matching these features of the investment rate distribution. This confirms the finding in Khan and Thomas (2008) that carefully chosen idiosyncratic productivity shocks can generate enough lumpiness without the need for nonconvex costs to be introduced in the model. The main remaining discrepancy between our model and the data is that we underpredict the asymmetry between positive and negative investments, likely due to not including growth in our model. 5. Quantitative results We consider two alternative types of reforms in both of which the corporate profits tax rate τcis permanently reduced and the government budget remains balanced. The two types of reforms differ in the tax instruments used in order to maintain the same level of long run revenue. In the first type of reform, both dividend and capital gains taxes are adjusted, whereas in the second only dividend taxes are adjusted. In both cases, we use labor taxes to balance the budget during the transition. For each type of reform, we discuss first a specific reform that reduces the corporate tax rate to zero. We discuss both the long run effects and the transitional, distributional and welfare effects of this case. Since transitional effects can be important for welfare, we also consider alternative assumptions regarding the extent to which a reform is anticipated in advance of its implementation. At the end, we also consider a range of values for the new level of τcto determine numerically the optimal level of corporate taxes. 5.1 Using equal dividend and capital gains taxes 5.1.1 Long run effects The first column of Table 7displays the long run effects of a reform that cuts corporate profits taxes to zero and replaces them with dividend and capital gains taxes, maintaining τd=τg. In the long run, the reform leads to a decrease in aggregate capital but TFP increases and this leads to an increase in aggregate output. These changes are a result of a combination of several counteracting mechanisms, which can be understood with reference to the proposition of Section 3.2.Itishelpfulto distinguish between mechanisms that affect all firms in a similar fashion, which in turn can be used to explain changes in aggregate capital, and mechanisms that have potentially opposite effects on different firms. The latter are used to explain changes in TFP, which arise from changes in the distribution of firms. Consider first the intuition for the decrease in aggregate capital. In the modified economy of the proposition in Section 3.2, the combined marginal tax rate τ≡τc+ τg(1−τc)on the return to capital is maintained fixed after the reform. This ensures that the optimal choices of firms and households at the margin remain the same and, 334 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) at the same time, the overall tax revenues for the government also remain the same. In contrast, in our benchmark economy, maintaining the same combined marginal tax rate would not ensure the same tax revenues for the government. This is because part of the adjustment costs are not deductible from corporate taxes (φ<1 as opposed to φ=1), but all adjustment costs are implicitly deducted from shareholder taxes, since dividends and capital gains are defined after payment of adjustment costs. As a result, switching from corporate taxes to shareholder taxes reduces the tax base and shareholder taxes have to rise to a point in which the combined tax rate τis higher than before (it increases from 47.2% before the reform to 50.6% after the reform). In turn, a higher marginal tax rate on the return to capital pushes investment and capital of all firms downwards. In addition to the effect through tax revenues, there is another effect that tends to reduce the incentives of firms to invest even if τwere to remain fixed. As reflected in the last term of equation (8), one of the benefits of increasing capital is that it lowers future adjustment costs. This benefit is taxed only partly by the corporate tax, but it is fully taxed under shareholder taxes. In other words, switching to shareholder taxes increases the marginal tax rate on this benefit and it also lowers the incentives to invest. Before moving on to the intuition regarding TFP changes, we briefly discuss the dependence of these results on the value of φ. It is clear from the preceding discussion that alowervalueofφwill lead to stronger effects on aggregate capital. In other words, the lower the value of φ, the larger the increase in the combined tax rate after the reform, and the larger the decrease in the aggregate capital stock. This is what we see in Table 7, Table 7. Eliminating corporate income taxes. τg=τdaτdb Reform ϕ=0.52 (Benchmark) ϕ=0.00 ϕ=1.00 ϕ=0.52 (Benchmark) Tax Rates τc0.0 0.0 0.0 0.0 τd50.6 52.4 47.2 43.7 τg50.6 52.4 47.2 20.0 Long Run Aggregates (% change) Y0.11.1−0.5 9.7 K−4.4 −6.4 −0.4 40.6 C−0.9 −0.6 −0.3 8.5 TFP 1.5 3.2 −0.4 −1.4 w0.11.1−0.5 9.7 r−2.0 −3.5 0.5 −2.2 Welfare (%)c Overall 0.39 1.08 0.01 −0.21 Aggregate 0.13 0.65 0.02 −0.84 Distributional 0.25 0.42 −0.01 0.64 aIn this reform, dividend and capital gains taxes change together, equalized to each other. bIn this reform, capital gains taxes are kept constant at their benchmark levels (τg=0.20). cSocial welfare gain/loss in consumption equivalent terms. It incorporates the effects of transition. Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 335 Table 8. Effects of eliminating τcon capital distribution across productivity levels. Productivity (z) z1z2z3z4z5z6z7z8z9z10 Change (%) in E(k|z)a Reform—(τcvs. τd=τg) Case—(ϕ=0.52) (Benchmark) −23.7 −20.5 −17.2 −13.7 −10.0 −5.9 −1.4 3.6 9.0 14.5 Case—(ϕ=0.00) −41.7 −36.4 −30.8 −24.7 −17.8 −9.8 −0.7 9.7 21.7 34.0 Case—(ϕ=1.00) 2.4 1.9 1.4 0.9 0.3 −0.2 −0.9 1.5 −2.2 −2.9 Reform—(τcvs. τd) Case—(ϕ=0.52) 66.0 60.2 55.2 50.4 45.8 41.3 36.9 32.9 29.7 27.6 aE(k|z)represents the mean capital conditional on productivity z. with the φ=0 case exhibiting the largest decrease in aggregate capital and the φ=1 case the smallest one. Consider now the intuition for changes in the distribution of capital across firms, and hence, TFP. There are two opposing forces. First, the proposition of Section 3.2 ensures that the reform is distributionally neutral by adding the term χjt ≡(qjtkj,t+1− qjt−1kjt )−(kj,t+1−kjt )to taxable corporate income Tjt. Since shareholder taxes implicitly tax the adjusted income ˜ Tjt =Tjt +χjt, this makes the tax base of corporate and shareholder taxes equivalent. Firms with relatively high productivity have relatively low values of χjt due to the fact that their investment rates are currently higher than the long run and their marginal qis falling, while the opposite is true for firms with relatively low marginal productivity. Thus, in the absence of this adjustment, a switch to shareholder taxes imposes relatively higher burden to unproductive firms with high χjt and a lower burden to those that are more productive and have a low χjt, leading to positive capital reallocation. Second, a reduction in the corporate tax rate essentially increases the effects of adjustment costs by virtue of shifting some of the burden of these costs away from the government and back to the firm. This increases the dispersion in marginal qand, therefore, the misallocation of capital due to adjustment costs. This second effect becomes stronger as the deductibility φof adjustment costs increases. As is evident in Table 7, the first effect dominates and TFP increases for the benchmark level of deductibility. For the case with φ=1, where adjustment costs are fully deductible, the second effect dominates and TFP decreases. These reallocation effects can also be seen in Table 8, which reports the average capital conditional on the value of zbefore and after the reform that eliminates corporate taxes. For φ=0.52 and φ=0, the effect of the reform is to reduce average capital for low-zfirms and increase it for high zfirms whereas the opposite is true for φ=1. 5.1.2 Transition, distribution, and welfare We use a standard utilitarian social welfare function to measure welfare and determine optimality. To better understand the welfare results, we use the method of Domeij and Heathcote (2004) to provide a decomposition of overall welfare into an aggregate and a distributional component. The aggregate component captures the effects of changes in aggregate consumption, both in the long run and along the transition, assuming these are proportionally distributed across individuals. The distributional component is computed as the residual, and thus captures any 336 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) departures from a proportional allocation of consumption effects. We also discuss how welfare effects differ by individual. The bottom panel of Table 7reports the welfare effects of the reform. The overall welfare gain is 0.39% in consumption equivalent terms. The decomposition into aggregate and distributional components indicates that there are both efficiency and distributional gains from the reform. The fundamental reason why the reform yields efficiency benefits is that it delivers higher production both in the short run and in the long run despite lower aggregate investment. This is due to the positive TFP effects, as is evident in the transition paths of macroeconomic aggregates displayed in Figure 1. Aggregate consumption exhibits a temporary but long lived (15 years) increase and a long run de- Figure 1. Transition paths when corporate income taxes are eliminated in reform (τcvs. τd=τg). φ=0.52. Note: The values for all variables are relative to their prereform levels. In all cases, the reform is announced at t=0 but the actual tax change occurs at t=0, 1, or 2 depending on the period of anticipation being 0, 1, or 2 years. Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 337 Figure 2. Individual welfare gains from eliminating corporate taxes in reform (τcvs. τd=τg). crease.23 Quantitatively, the transitional benefit dominates the long run cost and leads to a positive aggregate component of welfare of 0.13%. The reform also delivers distributional gains of 0.25% because high marginal utility households benefit and only a small fraction of low marginal utility households lose from the reform. This is illustrated inFigure 2, which plots the welfare gains and losses for each household (θ,)separately. Gains are decreasing in wealth and households with few or no stocks are the main beneficiaries, while only households with substantial wealth experience losses. The underlying reason has to do with the effects of the reforms on the after tax wage and after tax return. The after tax wage rises because of the increase in TFP, whereas the after tax return falls. As a result, households which earn primarily labor income tend to benefit whereas households that earn primarily capital income (i.e., high wealth, low marginal utility households) lose. Note that these distributional implications stand in sharp contrast to the findings in the literature regarding corporate tax cuts (e.g., Domeij and Heathcote (2004)) where such reforms are typically found to have negative distributional effects. The fundamental reason for this difference is the use of a capital tax (shareholder taxes in this case) to replace the corporate tax as opposed to using a labor tax. In existing literature, corporate tax revenues are made up using labor taxes and this implies that after tax wages drop as a result of the reform despite the positive general equilibrium effect on before tax wages. The bottom panel of Table 7also reports welfare effects for the cases φ=0andφ=1. Welfare gains are decreasing in the degree of deductibility of adjustment costs. In the ex- 23The long run decrease is an artifact of the simplifying assumption that adjustment costs are lost resources. This could be avoided by rebating costs back to the households at the cost of increasing computational time and introducing unintended distributional effects. Since the model yields a positive aggregate component of welfare despite this limitation, we have opted for not rebating the costs. We expect welfare gainstobelargerifthisisaddressed. 338 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) Table 9. Welfare effects with anticipation—elimination of corporate income taxes benchmark economy (ϕ=0.52). Years of Anticipation 0 1 2 Case τg=τd: Financing with Dividend and Capital Gains Taxes Welfare (%)a Overall 0.39 0.40 0.41 Aggregate 0.13 0.13 0.12 Distributional 0.25 0.27 0.29 Case τd: Financing with Dividend Taxes Welfare (%)a Overall −0.21 −1.03 −1.13 Aggregate −0.84 −1.20 −1.18 Distributional 0.64 0.18 0.05 aSocial welfare gain/loss in consumption equivalent terms. It incorporates the effects of transition. treme case of full deductibility welfare gains are small, which is not surprising given the small effects on long run aggregates in this case. However, the main message of this exercise, namely that eliminating corporate taxes in favor of shareholder taxes is welfare improving, remains true regardless of the value of φ. To the extent that adjustment costs reflect installation costs not fully deductible from the tax base, these gains can be significant. An important robustness check is to investigate whether the welfare effects are sensitive to the assumption that tax reforms are unanticipated. For this reason, we have also computed transitions and welfare under the assumption of anticipation, with the period of anticipation being 1 or two years. The welfare effects for these experiments are reported in Table 9. Anticipation does not make a significant quantitative difference for this type of reform. If anything, welfare gains rise with the length of anticipation period due to slightly better distributional effects. This is because the labor tax adjustment over the transition is smoother and implies a smaller, and more short lived, temporary drop in after tax wages. The overall conclusion is that this reform can deliver both efficiency and distributional gains and these positive aspects are robust to different anticipation periods. 5.1.3 Optimal corporate tax Although the elimination of corporate taxes in favor of shareholder taxes delivers welfare gains, τc=0 might not be the optimal choice. We investigate this by repeating the benchmark experiment for a range of different values of τcand determining numerically the choice of τcthat maximizes welfare. Figure 3shows the overall welfare gains, as well as the decomposition to aggregate and distributional components, for several values of τcfrom 0 up to the prereform value of 0.34. All cases considered yield positive overall welfare gains and these gains are increasing the larger the reduction in τc. This is true for both the aggregate and the distributional compo- Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 345 Table 11. Reforms with lower capital adjustment cost. Reform A: τcvs. τg=τdB: τcvs. τd Tax Rates τc0.00 0.00 τd0.52 0.47 τg0.52 0.20 Long Run Aggregates (% change) Y 0.14 3.34 K−5.19 37.10 C−0.94 3.28 TFP 1.82 −6.33 w 0.15 3.34 r−2.42 −0.12 in Cooper and Haltiwanger (2006) as opposed to the one in Compustat which we use in our benchmark. The volatility difference is substantial, the LRD suggests 0.337 whereas Compustat suggests 0.156. We chose to use the latter number because the LRD has plant level data for manufacturing firms, whereas Compustat has firm level data for publicly traded corporations, which is more relevant for the tax experiments we consider. However, given the substantial difference, we conduct a sensitivity experiment where we recalibrate the adjustment cost parameter ψto match the higher investment rate volatility. Table 11 reports the results and illustrates that this does not affect our results qualitatively but it does have a quantitative effect. The main difference is that the long run effects of the two reforms on TFP are now more pronounced. TFP increases by more in the τd=τgcase and it falls by more in the τd>τ gexperiment. In this sense, our benchmark results can be viewed as a lower bound. We have also experimented with introducing fixed adjustment costs similar to Khan and Thomas (2008) to our model. A small fixed cost can help increase the fraction of inactive firms (less than 1% investment rate) from 6% to 8% in our benchmark economy. However, the size of the cost needed is small and as a result, neither the other moments of the investment rate distribution in our benchmark nor the effects of the tax changes are substantially affected.28 6.2 Debt financing In our tax change experiments, financing a decrease in τcthrough increasing τdonly is found to be welfare reducing. To a large extent, this is because equity financing becomes costly and this generates capital misallocation. In practice, firms can switch to debt financing, and thus potentially mitigate the distortionary effects of the tax wedge, but our benchmark model abstracts from debt financing. In this section, we introduce debt financing and show that the presence of the debt financing option does not change our results significantly. The main reason is that these tax changes also reduce incentives for 28Numerical results are not included in the interest of space considerations. 346 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) debt financing, and thus there is little scope for substituting away from equity financing into debt financing in response to the tax changes. In contrast, we find some substitution from equity to debt in the case τd=τgbecause, in that case, the tax advantage of debt is actually increased. In the extended model, firm j’s financing constraint is djt +xjt +(xjt,kjt )+˜ rtbjt +ξ˜ rt b2 jt kjt =π(kjt,zjt;wt)−τcTjt +sjt +bjt+1−bjt, Tjt =π(kjt,zjt;wt)−δkjt −φ(xjt,kjt )−˜ rtbjt1+ξbjt kjt , where bjt denotes debt outstanding at t,˜ rtis the before tax interest on debt and ξ˜ rt b2 jt kjt represent costs of debt holding. Firms tradeoff the tax benefit of debt arising from its deductibility from corporate taxes against a marginal cost of debt that is increasing in the debt to capital ratio bjt kjt . Debt is held by households who obtain an after tax interest (1−τi)˜ rtwhere τiis the tax on interest income. In equilibrium, the after tax interest on debt is equalized to the after tax return on stocks rtand households are indifferent between the two assets. The firm’s optimal choice of debt is described by the Euler equation γjt =Et1+(1−τc)˜ rt+11+2ξbjt+1 kjt+1γjt+1 1+1−τi 1−τg ˜ rt+1 , where γjt =1−τd 1−τg+λd jt is the multiplier on firm j’s financing constraint. In the benchmark economy, where τd=τgand γjt =1, this gives a firm’s debt to capital ratio as bjt+1 kjt+1 = 1−τi (1−τg)(1−τc)−1 2ξ reflecting the tradeoff between the tax advantage and the marginal cost of being in debt that is increasing in bjt+1 kjt+1. The tax advantage obtains whenever the combined marginal tax rate of corporate income τ≡τc+τg(1−τc)is larger than the tax rate on debt interest income τi. When equity issuance becomes costly due to the reform making τd>τ g, firms which need external financing (γjt >E tγjt+1) will choose higher debt to capital ratios than firms which do not. The long run results of the two reforms where τcisreducedtozerowhilerevenueis made up with τd=τgor with τdonly keeping τgfixed are shown in Table 12.29 29For the quantitative results, the debt cost parameter ξis calibrated to match a ratio of debt to assets of 0.12 based on Hennessy and Whited’s (2007) reported number from Compustat data and τi=τl=0.28. Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 347 Table 12. Reforms in the presence of debt financing.a Reform A: τcvs. τg=τdB: τcvs. τd Tax Rates τc0.00 0.00 τd0.52 0.42 τg0.52 0.20 Long Run Aggregates (% change) Y−0.3 9.2 K−5.4 36.3 C−1.7 8.4 TFP 1.5 −0.8 w−0.3 9.2 r−2.8 −0.8 Financial Aggregates (% change) S−55.8 −98.9 Debt Issuance 56.7 −23.7 B33.7−83.8 aDebt issuance is calculated as the aggregate of bj,t+1−bj,tover all firms jwith bj,t+1−bj,t>0. In the first case (τd=τg), the combined marginal tax rate on corporate income increases relative to the benchmark (from τ=0.472 to τ=0.52), and thus the tax incentive for holding debt is stronger. The debt to capital ratio increases from 0.15 to 0.21 and some of the equity issuance is replaced by increased debt issuance. In terms of the macroeconomic aggregates, the results are similar to the case of no debt. The second case (τd>τ g) is the one that is more interesting. In order to finance a move to a zero corporate tax, the dividend tax rate has to be significantly increased. In turn, this creates a large tax wedge (τdτg), which makes equity very costly and induces almost all growing firms to grow internally instead of using equity financing. This was the case in our benchmark experiments without debt and, as reflected in Table 12, this is still the case in the presence of debt. Despite that, we do not find that debt financing is used to replace equity financing. Instead, total debt held falls by more than 80% and even aggregate new debt issuance (for firms that issue positive amounts) decreases by 24%. The reason is that the combined marginal tax rate on corporate income is significantly reduced (from τ=0.472 to τ=0.20) and this has made the tax advantage of debt disappear. The majority of firms hold zero debt because tax considerations make holding any debt suboptimal. Only very few severely cash strapped (γjt Etγjt+1)firmsare willing to issue new debt. As a result, it is still the case that external financing becomes much less prevalent and the effects of this reform on the macroeconomic aggregates are similar to the case of no debt. 7. Conclusion We find that reforms which replace corporate income taxes with shareholder taxes can enjoy widespread popular support. Whether such reforms are efficiency enhancing 348 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) hinges critically on the mix of dividend and capital gains taxes and the degree to which the new policy is anticipated. In particular, if dividends and capital gains are taxed at the same rate, we predict efficiency gains that are robust to different degrees of anticipation. However, when the shareholder tax used is only the dividend tax rate, this policy can have negative consequences on efficiency because the resulting misallocation of capital creates more distortions than the ones solved by the removal of the corporate tax. Moreover, this reform can have additional negative consequences to the extent that it is anticipated. These results are derived in a rich environment, consistent with key features of the US economy, such as wealth inequality at the household level, imperfect risk sharing, productivity differences across firms, and endogenous financing decisions for the corporate sector. All of these components are important in evaluating the effects of different types of capital income taxes. While incorporating those components that are crucial for the policy question at hand, we have abstracted from several other potentially important margins. First, the literature has identified additional channels through which a corporate profits tax cut can affect macroeconomic performance, which we did not incorporate in our model. These include the choice of legal form of organization, the effects on employment as well as the possibility of international capital flows. It is noteworthy that studies which include these other channels seem to reach a similar conclusion to ours, namely that a reduction in corporate profits taxes can be beneficial to the economy. This paper contributes to the discussion by suggesting an alternative way of financing this tax cut that can increase popular support for such a reform. Second, our dynamic general equilibrium model does not incorporate some of the intricacies of the actual tax code that could play a role in policy decisions. For example, a significant fraction of household savings are not subject to dividend or capital gains taxes because they are held in retirement accounts and that is not captured in the model. More importantly, we have not modelled capital gains taxes on a realization basis so our model cannot capture potentially relevant aspects of the taxation of capital gains such as the timing of realizations and lockin effects, the deferral of realization until death to take advantage of tax forgiveness at death as well as the issues arising from using a nominal basis for capital gains in the actual tax code. We share this limitation with the vast majority of the existing literature. This is especially the case for work on general equilibrium macroeconomic models like ours, not only because of the complications introduced by the timing of the realization decision but also due to the need to keep track of past purchase prices for each component of an agent’s portfolio. This is an important modeling question to be addressed in future research. Third, consistently with most of the literature, we do not provide a model of the choice between dividends and share buybacks as a means to distribute profits. In our model, when dividend and capital gains taxes are equalized the choice is irrelevant. However, whenever dividend taxes are higher than capital gains taxes, firms would prefer to use buybacks and our model simply assumes this option away. This tax arbitrage is another reason why equalization of dividend and capital gains taxes is often proposed in practice and that margin is missing in our model. On the other hand, firms’ ability Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 349 to use buybacks might reduce the importance of the financing frictions introduced by dividend taxes. A careful model of the tradeoffs in using dividends versus buybacks as a means to distribute profits is beyond the scope of the present study. With these caveats in mind, our main conclusion is that a tax code that focuses more on taxing shareholders directly rather than indirectly through corporations would be a step in the right direction, in the sense that a large majority of households could agree with and benefit from it. Appendix A: Proof of Proposition 1 The goal is to show that all long run equilibrium conditions are satisfied for the new taxes τ∗ c,τ∗ sand dividend payout (djt −sjt )∗but for otherwise identical allocations and prices to the ones before the reform. We focus only on the conditions that involve the taxes and dividend payout, since the rest are trivially satisfied. Throughout the section, we let πjt ≡π(kjt,zjt;wt). Firms’ conditions have to be adjusted according to the new tax code assumptions. Using the newly defined taxable corporate income in (11), the firms’ financing constraint reads djt −sjt =πjt −(xjt,kjt )−xjt −τc˜ Tjt. After the reform this financing constraint is satisfied by construction of the dividend payout specified in the proposition. Recall that with equal capital gains and dividend taxes, λd t=λs t=0 and note that we use τsto denote both dividend and capital gains tax rates since they are equal. The first-order condition for investment is now qjt =1+(1−τcφ)x(xjt,kjt )−τcφ(1−qjt ). After some rearrangement, this gives qjt =1+x(xjt,kjt ) which is still satisfied after the reform for the same allocations since no tax term is involved. The capital first-order condition is now qjt =1 1+r 1−τs Et(1−δ)qj,t+1+(1−τc)∂πjt+1 ∂kjt+1 +1 1+r 1−τs Etτcqjt −(1−δ)qjt+1−(1−τcφ)k(xjt+1,kjt+1). After some manipulation, this can be simplified to r=1 qj,t Et(1−τs)(1−τc)(1−δ)qj,t+1−qj,t+∂πjt+1 ∂kjt+1−(1−τcφ)k(xjt+1,kjt+1). It is easy to see that if φ=1, the previous condition is also still satisfied for the same allocation when the overall tax wedge (1−τc)(1−τs)is kept fixed. 350 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) The household budget constraint and the first-order condition for stocks are the same as in Section 2. At steady state, these are cit +P(θit −θit−1)=(1−τl)wit +(1−τs)(D−S)θit−1 and r≡(1−τs)(D−S) P. From the households’ perspective, taxes may affect the overall payoff (1−τs)(D−S) on the right-hand side of both of those conditions. Using the financing constraint of a firm together with the taxable income in equation (10) and aggregating over all firms j, this term can be written as (1−τs)(D−S) =(1−τs)−−X−τc˜ Tjt dj =(1−τs)−−X−τc−−X+(qjtkj,t+1−qjt−1kjt ) =(1−τs)(1−τc)(−−X)−τc(1−τs)(qjtkj,t+1−qjt−1kjt ), where =πjtdj,=(xjt,kjt )dj and we have used φ=1 again. In a stationary distribution, the last term disappears. As a result, every household’s budget constraint (1) and Euler equation (2) are still satisfied after the reform. It follows that government revenues are the same after the reform, and thus the government’s budget is also satisfied. This completes the proof. Appendix B: Modeling depreciation allowances In this section, we show how the present value of depreciation allowances can be captured through the parameter φ. To model depreciation allowances, we closely follow Auerbach (1989). Throughout the section, we let πjt ≡π(kjt,zjt;wt). Let Gjt represent the depreciation allowances at time tarising from all past capital expenditures including installation costs The constraints of the firm can be written as djt +xjt +(xjt,kjt )=πjt −τc[πjt −Gjt]+sjt, kjt+1=(1−δ)kjt +xjt, Gjt = t−1  u=−∞ δ(1−δ)t−1−uxju +(xju,kju), njt ≥0, djt ≥0. Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 351 Using the capital accumulation equation to express kjt in terms of all past investment as kjt = t−1  u=−∞ (1−δ)t−1−uxju Gjt can equivalently be written as Gjt =δkjt + t−1  u=−∞ δ(1−δ)t−1−u(xju,kju). This makes explicit the fact that total allowances are composed of the standard depreciation term δkjt plus a second component corresponding to the “depreciation” of installation costs. To simplify this second component, let the discount factor of the firm between periods tand sbe denoted by Mt,s=(s−t n=11 1+rt+n 1−τg ). If we write the Lagrangian of the firm’s problem and assume that the multiplier on the financing constraint is equal to M0tγjt, the term involving Gjt canbewrittenas ∞  t=0 M0,tγjtτcGjt =τc ∞  t=0 M0tγjtδkjt +δ t−1  u=−∞ (1−δ)t−1−u(xju,kju) =τc ∞  t=0 M0,tγjtδkjt +jt(xjt,kjt ), where jt =δ ∞  s=1 Mt,t+s γjs+t γjt (1−δ)s−1 and we have used the fact that M0,s+t=M0tMt,t+s. This has collected together all the future depreciation allowances arising from the time tinstallation costs and expressed them in present value terms. The total fraction of the current installation costs (xjt,kjt ) that is ultimately deducted is, in present value terms, represented by jt. Using this expression, the financing constraint of the firm can be written as djt +xjt +(xjt,kjt )=πjt −τcπjt −δkjt −jt(xjt,kjt )+sjt which essentially implies that the firm is deducting a fraction jt of capital adjustment costs every period. Because of the presence of time varying endogenous variables in the infinite sum of jt, a full numerical implementation of this problem would require an additional state variable, essentially capturing the “stock” of installation costs paid.30 Given the additional computational complexity this would introduce, we instead choose to follow Auerbach’s approach, which is to simply compute the value of at the long run equilibrium. We focus on the case τd=τgwhere γjs+t=γjt. Replacing the long run value of the 30A recursive formulation can be provided upon request. 352 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) firm’s discount factor, the value of is equal to =δ r 1−τg +δ . References Abel, A. B. (1983), “Tax neutrality in the presence of adjustment costs.” The Quarterly Journal of Economics, 98 (4), 705–712. [327] Abraham, A. and E. Carceles-Poveda (2010), “Endogenous trading constraints with incomplete asset markets.” Journal of Economic Theory, 145, 974–1004. [329,330] Anagnostopoulos, A., O. E. Atesagaoglu, E. Cárceles-Poveda (2022), “Supplement to ‘Financing corporate tax cuts with shareholder taxes’.” Quantitative Economics Supplemental Material, 13, https://doi.org/10.3982/QE1167.[324] Anagnostopoulos, A., E. Carceles-Poveda, and D. Lin (2012), “Dividend and capital gains taxation under incomplete markets.” Journal of Monetary Economics, 59 (7), 898–919. [316,319,328,340] Atesagaoglu, O. E. (2012), “Taxes, regulations and the corporate debt market.” International Economic Review, 53 (3), 979–1004. [316,330] Auerbach, A. J. (1989), “Tax reform and adjustment costs: The impact on investment and market value.” International Economic Review, 30 (4), 939–962. [331,350] Cárceles-Poveda, E. and D. Coen-Pirani (2010), “Owning capital or being shareholders: An equivalence result under incomplete markets.” Review of Economic Dynamics, 13, 537–558. [325] Castañeda, A., J. Diaz-Gimenez, and J. V. Ríos-Rull (2003), “Accounting for earnings and wealth inequality.” Journal of Political Economy, 111 (4), 818–857. [329] Chen, D., S. Qi, and D. E. Schlagenhauf (2018), “Corporate income tax, legal form of organization, and employment.” American Economic Journal: Macroeconomics, 10 (4), 270– 304. [319] Conesa, J. C. and B. Dominguez (2013), “Intangible capital and Ramsey capital taxation.” Journal of Monetary Economics, 60, 983–995. [319] Conesa, J. C., S. Kitao, and D. Krueger (2009), “Taxing capital? Not a bad idea after all!” American Economic Review, 99 (1), 25–48. [316] Cooper, R. and J. Haltiwanger (2006), “On the nature of adjustment costs.” The Review of Economic Studies, 73 (3), 611–633. [330,332,345] Dammon, R. M., C. S. Spatt, and H. H. Zhang (2001), “Optimal consumption and investment with capital gains taxes.” The Review of Financial Studies, 14 (3), 583–616. [320] Quantitative Economics 13 (2022) Financing corporate tax cuts with shareholder taxes 353 Dammon, R. M., C. S. Spatt, and H. H. Zhang (2004), “Optimal asset location and allocation with taxable and tax-deferred investing.” Journal of Finance, 59 (3), 999–1037. [320] Davila, J., J. H. Hong, P. Krusell, and J. V. Ríos-Rull (2012), “Constrained efficiency in the neoclassical growth model with uninsurable idiosyncratic shocks.” Econometrica,80(6), 2431–2467. [328] Diaz, A., J. Pijoan-Mas, and J. V. Ríos-Rull (2003), “Habit formation: Implications for the wealth distribution.” Journal of Monetary Economics, 50 (6), 1257–1291. [329] Diaz-Gimenez, J., A. Glover, and J. V. Rios-Rull (2011), “Facts on the distributions of earnings, income, and wealth in the United States: 2007 update.” Federal Reserve Bank of Minneapolis Quarterly Review, 34 (1), 2–31. [329] Domeij, D. and J. Heathcote (2004), “On the distributional effects of reducing capital taxes.” International Economic Review, 45 (2), 523–554. [316,330,335,337] Favilukis, J., S. C. Ludvigson, and S. van Nieuwerburgh (2017), “The macroeconomic effects of housing wealth, housing finance, and limited risk-sharing in general equilibrium.” Journal of Political Economy, 125 (1), 140–223. [319,322] Feenberg, D. and E. Coutts (1993), “An introduction to the taxsim model.” Journal of Policy Analysis and Management, 12 (1), 189–194. [330] Fehr, H., S. Jokisch, A. Kambhampati, and L. J. Kotlikoff (2013), “Simulating the elimination of the U.S. corporate income tax.” NBER working Paper No. 19757. [319] Gavin, W. T., F. E. Kydland, and M. R. Pakko (2007), “Monetary policy, taxes, and the business cycle.” Journal of Monetary Economics, 54 (6), 587–1611. [320] Gourio, F. and J. Miao (2010), “Firm heterogeneity and the long run effects of dividend tax reform.” American Economic Journal: Macroeconomics, 2 (1), 131–168. [316,319,321, 322,328,330,331,332,341] Gourio, F. and J. Miao (2011), “Transitional dynamics of dividend and capital gains tax cuts.” Review of Economic Dynamics, 14 (2), 368–383. [342] Hennessy, C. A. and T. M. Whited (2007), “How costly is external financing? Evidence from a structural estimation.” The Journal of Finance, 62 (4), 1705–1745. [346] Khan, A. and J. Thomas (2008), “Idiosyncratic shocks and the role of nonconvexities in plant and aggregate investment dynamics.” Econometrica, 76 (2), 395–436. [333,345] McGrattan, E. R. and E. C. Prescott (2005), “Taxes, regulations, and the value of US and UK corporations.” Review of Economic Studies, 72, 767–796. [316,325] Mendoza, E., A. Razin, and L. Tesar (1994), “Effective tax rates in macroeconomics. Cross country estimates of tax rates on factor income and consumption.” Journal of Monetary Economics, 34 (3), 297–323. [330] Miao, J. and P. Wang (2014), “Lumpy investment and corporate tax policy.” Journal of Money, Credit and Banking, 46 (6), 1171–1203. [319] 354 Anagnostopoulos, Atesagaoglu, and Cárceles-Poveda Quantitative Economics 13 (2022) Santoro, M. and C. Wei (2011), “Taxation, investment and asset pricing.” Review of Economic Dynamics, 14 (3), 443–454. [316] Tauchen, G. and R. Hussey (1991), “Quadrature-based methods for obtaining approximate solutions to nonlinear asset pricing models.” Econometrica, 59 (2), 371–396. [330] Co-editor Christopher Taber handled this manuscript. Manuscript received 5 July, 2018; final version accepted 5 July, 2021; available online 2 September, 2021.