GB792 Unit 9 Research Design Proposal Example

When State Tax Rates Move: A Unit 9 Research Design Proposal on Taxes and Corporate Borrowing

[Student Name]

Purdue University Global

GB792: Advanced Topics in Finance

Unit 9 Research Design Proposal

[Instructor Name]

[Date]

The proposed study has not been carried out; the data sources, methods and prior studies cited are real.

Introduction and Lineage

Whether taxes affect how much firms borrow is one of the oldest questions in corporate finance. Interest is deductible, so a higher tax rate makes debt more valuable, and trade-off theory predicts that firms will borrow more when rates rise. Testing that prediction is difficult, because tax rates rarely change on their own: a change in the federal rate affects every firm at once, along with whatever else changed in the economy that year.

State corporate tax changes offer a way around that difficulty. Because states change their rates at different times, firms in states that raised rates can be compared with similar firms in states that did not. Ljungqvist et al. (2017) used this staggered variation to study how taxes affect corporate risk taking, showing that state tax changes can identify the effect of taxes on corporate choices. Doidge and Dyck (2015), studying a surprise decision by Canada to tax income trusts, found that tax shields added measurably to firm value and that firms adjusted their debt, payouts and cash in response to the change. Whether firms respond to tax increases and tax cuts with equal force has received less attention.

This proposal asks whether American firms change their borrowing after state tax changes in the period since 2009, using data a candidate can obtain without a commercial subscription, and whether increases and cuts produce responses of the same size.

Hypotheses

The first hypothesis is that firms headquartered in a state that raises its corporate tax rate increase their debt ratios in the following years, relative to firms in states with no change. The second is that firms in states that cut their rate reduce their debt ratios by a smaller amount, an asymmetry that would sit uneasily with a simple trade-off. The third, a test of mechanism, is that the response to increases is larger for firms with more taxable income, since only firms that pay tax benefit from a larger deduction.

Data

Financial statements. The SEC publishes Financial Statement Data Sets, extracted from the XBRL data that public companies have attached to their filings since 2009. They contain balance sheet and income statement items for every filer, identified by standard tags. The debt ratio will be computed from total debt, using the tags for long-term debt, current portion and short-term borrowing, and total assets.

Headquarters state. Each EDGAR filing header reports the company's business address. The state of that address will be used as the headquarters state, recorded for each fiscal year so that relocations are captured.

State tax rates. The top marginal corporate income tax rate for each state and year will be taken from a published compilation, such as the annual state tax tables issued by a tax research organization, and each change will be checked against the state's enacting statute to confirm the effective date.

Estimator

The design is a staggered difference-in-differences. Firms in states that change their rate are treated from the year of the change; firms in states without a change in a given window serve as controls. Because states change rates at different times, a conventional two-way fixed effects regression would compare later-treated firms with earlier-treated ones as if the latter were controls, which can bias the estimate when effects grow over time. Goodman-Bacon (2021) showed how this bias arises. The proposal therefore uses the estimator of Callaway and Sant'Anna (2021), which compares each group of treated firms only with firms not yet treated or never treated, and then aggregates the results.

The analysis will be presented in event time, with leads and lags from three years before to four years after each change. The leads test whether debt ratios were already moving before the change, which would suggest anticipation or a confounding trend. Tax increases and tax cuts will be estimated separately to test the asymmetry.

The leads before each tax change are the design's own warning light: if borrowing moves before the rate does, something other than the tax is at work.

Threats and Responses

Headquarters is not where taxes are paid. Firms pay state tax in many states, apportioned largely by where they make sales. A firm headquartered in a state with a rate increase may pay little tax there. The response is to treat headquarters as an imperfect measure of exposure, which biases estimates toward zero, and to test whether effects are larger for firms whose subsidiaries, listed in exhibit 21 of the annual report, are concentrated in the headquarters state.

States raise taxes when their economies change. A state may raise its rate during a budget crisis, when local firms are also under pressure. The response is to include state-level economic controls, such as unemployment and income growth, and to compare firms in the same industry across states.

Anticipation. Tax increases are often debated before enactment. The event-time leads will show whether firms adjusted early, and enactment dates will be recorded alongside effective dates.

Relocation. Firms might move headquarters in response to tax changes. Relocations will be identified from filing headers and examined separately.

Feasibility

The proposal depends on the data being usable. A table will record, for each variable, its source, its XBRL tag and the share of firm-years where it is missing. Because companies tag their borrowing in several different ways, a rule for combining the debt tags will be written and tested on one year of filings before the main analysis. State rate histories are short documents that can be checked by hand. The full dataset can be assembled with free software in a few weeks.

What the Study Could Show

If firms borrow more after tax increases but reduce debt little after cuts, the study would document an asymmetry that trade-off theory does not predict and that later work would need to explain. If they respond to both in similar measure, the evidence would fit the trade-off view. If they respond to neither, the result would suggest that, at least for state taxes and the firms filing XBRL data, tax incentives play a smaller role in borrowing than trade-off theory predicts.

References

Callaway, B., & Sant'Anna, P. H. C. (2021). Difference-in-differences with multiple time periods. Journal of Econometrics, 225(2), 200-230. https://doi.org/10.1016/j.jeconom.2020.12.001

Doidge, C., & Dyck, A. (2015). Taxes and corporate policies: Evidence from a quasi natural experiment. The Journal of Finance, 70(1), 45-89. https://doi.org/10.1111/jofi.12101

Goodman-Bacon, A. (2021). Difference-in-differences with variation in treatment timing. Journal of Econometrics, 225(2), 254-277. https://doi.org/10.1016/j.jeconom.2021.03.014

Ljungqvist, A., Zhang, L., & Zuo, L. (2017). Sharing risk with the government: How taxes affect corporate risk taking. Journal of Accounting Research, 55(3), 669-707. https://doi.org/10.1111/1475-679X.12157