Small Business Incorporation and Investment: The Role of Corporation Tax Michael Devereux and Li Liu Centre for Business Taxation, University of Oxford Public Economics UK, University of Bristol May 7-8, 2014
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
1 / 25
Motivation:Why Do Small Businesses Incorporate?
Limited liability?
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
2 / 25
Motivation:Why Do Small Businesses Incorporate?
Limited liability? Separation of ownership and control?
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
2 / 25
Number of Businesses in the UK, By Employment
Companies
Partnerships
Sole proprietorships
0
1,000,000 0 employees 10-99 employees
Devereux and Liu (Oxford)
Small Business Incorporation
2,000,000
3,000,000
1-9 employees 100 or more employees 8/5
3 / 25
Why Do Small Businesses Incorporate?
Limited liability? Separation of ownership and control? Better access to external finance for more investment? Information for lenders better if incorporated, and improves over time Lower tax
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
4 / 25
What is the Role of Corporation Tax?
Lower CT induces firms to incorporate conditional on personal tax Lower CT increases post-tax cash flow for companies and allows financially-constrained companies to invest more Lower CT reduces cost of capital for companies and provides greater incentive to invest
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
5 / 25
Our empirical approach
Use population of UK corporation tax records in 2001/02 - 2008/09 so data only on corporations, not unincorporated businesses Explore the UK 2006 abolition of zero starting rate as a quasi-natural experiment examine effect on incorporation, and effect on investment, via cost of capital and cash flow
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
6 / 25
Results
Estimate a significant positive effect of tax savings to incorporation On average, a 1 percentage point increase in tax saving raises probability of incorporation by 2% to 4%
Estimate effects of CT on investment by companies through 2 channels: a higher user cost of capital, and a reduction in internal cash flow available for investment
Effects of cash flow diminish with company age
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
7 / 25
Some Existing Literature
Corporate Investment and Financial Constraints Theoretical models of capital markets with asymmetric information: e. g. Stiglitz and Weiss (1981), and Myers and Majluf (1984) Empirical literature: Fazzari, Hubbard and Peterson (1988), Bond and Meghir (1994), Bond and Van Reenen (2007)
Taxes, Risk Taking and Entrepreneurship Mackie-Mason and Gordon (2002), Cullen and Gordon (2007)
Taxes and the Choice of Organizational Form Goolsbee (2004), de Mooij and Nicodeme (2008), Da Rin et al. (2011), Elschner (2013) and Liu (2014)
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
8 / 25
Basic Setup with Corporation Tax C oC
T +
C o C e x t , uninc
T+
C o C ext
T +
C o C int
MP
c a s h fl o w i n te r n a l
Devereux and Liu (Oxford)
I
A
A
I
e x te r n a l
Small Business Incorporation
8/5
9 / 25
Following a Decrease in the Corporation Tax C oC
T +
C o C e x t , uninc
T + ,n e w
C o C ext
T +
C o C int
MP
T + ,n e w
A
C o C int
i nt e r n a l H i g h e r n e t c a s h fl o w Devereux and Liu (Oxford)
I
e xte r n a l
A
I
e x t e r n al Small Business Incorporation
8/5
10 / 25
Data
Population of UK corporate tax returns (CT600), 2001/02 to 2008/09 10.7 million observations for 2.5 million companies Detailed and precise information on taxable profits and how they are determined
Around 90% of the tax records matched with company accounts in FAME Year of Incorporation: FAME Investment: qualifying expenditure on machinery and plant, including qualifying expenditure for FYA (CT600, Box 118), long-life assets and integral features (Box 120), and other machinery and plant (Box 121).
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
11 / 25
0
.1
.2
.3
.4
.5
Policy Reform to the Zero Starting Rate: Marginal Tax Rate in 2002/03
0
20k
40k
60k 80k Taxable Income
Retained Earnings Devereux and Liu (Oxford)
Small Business Incorporation
100k
120k
140k
Self-Employment Income 8/5
12 / 25
0
.1
.2
.3
.4
Zero Starting Rate Abolished in 2006/07
0
20k
40k
60k 80k Taxable Income
Retained Earnings Devereux and Liu (Oxford)
Small Business Incorporation
100k
120k
140k
Self-Employment Income 8/5
13 / 25
Tax Gains to Incorporate: Retained Profit Tax Gains to Incorporate (%): Average Tax RateSelf-Employment Income Average Tax RateCorporate Profit
Tax Gains to Incorporate: Retained Profit
20
10
0
-10
-20 0 Devereux and Liu (Oxford)
20
40 60 Corporate Taxable Profit (£1,000) Small Business Incorporation
80
100 8/5
14 / 25
Number of Newly Incorporated Firms
Number of New Companies
30000
20000
10000
0 0
20
40 60 Corporate Taxable Profit (£1,000) 2002/03-2003/04
Devereux and Liu (Oxford)
Small Business Incorporation
80
100
2006/07-2007/08 8/5
15 / 25
The Causal Effect of Tax Incentives on Incorporation
E (cit jTax_Gainit, Xit ) = exp (γi + λt + βtax Tax_Gainit + Xit βx ) cit : number of newly incorporated businesses in income bin i of £100 at time t γi : income bin dummies λt : time dummies Tax_Gainit : the difference between average tax rate for observed corporate profit, and the corresponding average tax rate if the corporate profit were earned as self-employment income Xit : other observed firm characteristics that may matter for incorporation Associated error term can be additive or multiplicative depending on functional form
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
16 / 25
Baseline Specification: Model Comparison
Estimation Model Tax Gains: Retained Profits
Income Bin FE Year FE No. of Observations No. of Income Bins
Devereux and Liu (Oxford)
Log Linear
Poisson GLM
Negative Binomial
Poisson Pseudo-MLE
0.043*** (0.001)
0.042*** (0.004)
0.046*** (0.001)
0.042*** (0.002)
Y Y 6,000 1,000
Y Y 6,000 1,000
Y Y 6,000 1,000
Y Y 6,000 1,000
Small Business Incorporation
8/5
17 / 25
Full Model Specification
Coefficients lower (1.9% to 2.2%) if based on tax saving for distributed dividends Effects robust to inclusion of covariates. such as turnover, total asset, and number of workers exclusion of the bunching region industry-level estimation controlling for industry fixed effects, industry-specific time trend, and industry-level covariates heterogeneous responses across different industries
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
18 / 25
Investment Equation
Static Specification: Iit Ki,t
1
= ∆ ln Yit
σ∆ ln CoCit + di + η t + εit
Iit Ki,t
: investment undertaken by company i in year t, scaled by 1 beginning-of-year book value of tangible asset Kt 1 ; approximation for ∆ ln Kit + δi
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
19 / 25
Excess Sensitivity to Cash Flow
Iit Ki,t
1
= ∆ ln Yit
σ∆ ln CoCit
γtax
∆Taxi,t Ki,t 1
1
+ di + η t + εit
∆Taxi,t 1 : one-year lagged increase in tax bill after abolishing the zero starting rate; =∆τ i,t 1 Taxable Profiti,t 1 ,where t = 2007 or 2008 Instrumented with ∆τ i,t 1 , change in the statutory marginal tax rate due to exogenous changes in the tax system Allow γtax to depend on "age" ie. time since incorporation
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
20 / 25
.224
.226
Cost of Capital .228 .23
.232
.234
Changes in the Cost of Capital
0
20
Devereux and Liu (Oxford)
40
60 80 100 Corporate Taxable Profit (£K) 2002/03
Small Business Incorporation
2006/07
120
140
8/5
21 / 25
0
Increase in Tax Liability (£) 500 1000 1500
2000
Increase in Tax Liability 2006/07, £000
0
Devereux and Liu (Oxford)
50Small Business Incorporation
100
150
8/5
22 / 25
Excess Sensitivity to Cash Flow: Empirical Findings
∆ ln Yit ∆ ln CoCit ∆Taxi,t Ki,t 1
1
∆Taxi,t Ki,t 1
1
Static Specification (1) (2) (3) 0.125*** 0.139*** 0.138*** (0.002) (0.003) (0.003) -1.895*** -1.696*** -1.677*** (0.066) (0.084) (0.086) -0.783*** (0.260)
Ageit
-1.371*** (0.229) 0.090*** (0.035)
Firm FE Year FE No. Obs
Devereux and Liu (Oxford)
Y Y 561,486
Y Y 381,499
Small Business Incorporation
Y Y 381,499
8/5
23 / 25
Excess Sensitivity to Cash Flow: Empirical Findings Error-Correction Specification (4) (5) (6) 0.354*** 0.354*** 0.352*** (0.003) (0.004) (0.004) -1.052*** -0.911*** -0.904*** (0.060) (0.070) (0.071) -0.481*** -0.519*** -0.516*** (0.003) (0.006) (0.007)
∆ ln Yit ∆ ln CoCit ECit 1 ∆Taxi,t Ki,t 1 ∆Taxi,t Ki,t 1
1
-0.352*** (0.211)
1
Ageit
-0.725*** (0.178) 0.057* (0.029)
Firm FE Year FE No. Obs Devereux and Liu (Oxford)
Y Y 526,457
Y Y 381,499
Small Business Incorporation
Y Y 381,499 8/5
24 / 25
Conclusions
Evidence that incorporation decisions depend on scale of tax advantage investment of companies affected by exogenous change to cash flow cash flow effect greater for younger companies, consistent with cost of external finance falling as banks have access to more information
Devereux and Liu (Oxford)
Small Business Incorporation
8/5
25 / 25