Sheppard on Earnings Stripping

Our favorite tax fashionista, Lee Sheppard, has an interesting article out today. See "News Analysis: The Fashion in Interest Deduction Restrictions," Tax Notes, Nov. 28, 2011, p. 1061.

Earnings Stripping

What is earnings stripping? First, some background. Many jurisdictions, including the United States, permit taxpayers to deduct interest expense in computing net taxable income. Deductible interest reduces pre-tax income and taxes owed to the revenue authority. The deductibility of interest expense creates a difficult challenge for policy makers. The issue will be lurking below the surface as Congress begins wading into the waters of tax reform.

Deductible interest creates a "bias" for debt-heavy capital structures ("overleverage"). The deduction for interest expense artificially decreases the "real" cost of debt relative to the "real" cost of equity. Overleveraged capital structures are more likely to capsize during the "trough" of business cycles, creating turmoil and trauma for stakeholders of the overleveraged business (employees, customers, local communities, etc.).

Deductible interest also permits business owners to "strip" taxable income out of the the tax base. This "earnings stripping" or "base erosion" involves variations on a theme. Sheppard's article focuses on cross-border earnings stripping.

A corporate parent with a foreign subsidiary may be able to fund the subsidiary with "shareholder loan" capital that pays interest. With proper treaty planning, (i) the shareholder loan pays interest, (ii) the interest expense is deductible, reducing taxable income and tax liabilities in the subsidiary's jurisdiction, and (iii) low or no withholding taxes apply to the interest payment. This type of structure permits the parent to "strip" earnings out of a higher-tax jurisdiction into a lower-tax jurisdiction.

Designing an "interest strip" is not rocket science. Tax advisors can use the basic structure in domestic transactions. Assume a group of pension funds pools capital and purchases a U.S. business organized as a corporation. The pension funds capitalize an acquisition vehicle with a mix of equity capital and "shareholder loan" capital. The decision to capitalize with equity and debt is effectively "deemed" to have economic substance. The target business generates taxable income after the acquisition. But the acquisition vehicle pays deductible interest expense on its shareholder loans. The pension funds are tax-exempt entities, so pay no tax when they accrue or receive interest. Voila, an "interest strip" -- shifting tax revenue from the U.S. Treasury to the pension funds.

(In the long run, query whether foreign and domestic pension funds will be able to use their tax-advantaged status to purchase every commercial enterprise on Earth. Marx would be vindicated. The working class -- through pension funds and their investment managers -- would control the means of production.)

U.S. Rules

Congress understands that earnings stripping can be a big problem. It enacted Code section 163(j) to address the issue. Sheppard snorts at the U.S. rules: "[I]t remains ridiculously easy to strip [income out of the United States] using interest deductions, despite a statute squarely aimed at foreign-parented groups."

I won't go into section 163(j) in detail. It limits a corporation's related-party interest deductions if the corporation has too much leverage and too much interest expense. It works in some cases, but has a couple obvious shortcomings. First, it doesn't apply to my domestic example above (where a group of pension funds acquire a U.S. business and use shareholder loans to interest strip). The statute only bites if the shareholder loan capital is provided by a "related" person.

Second, the U.S. limitation does not expressly contain any transfer pricing limitation. Take a technology business. Many technology businesses are capitalized with low ratios of third-party debt to equity ("leverage ratios"). But let's assume that a foreign parent corporation purchases a U.S. target corporation and injects leverage (a shareholder loan) into the target's capital structure. Assume that external leverage ratios in the industry are very low (10% or lower). But the foreign parent capitalizes its new subsidiary with $20x of equity capital and $80x of shareholder loan capital (an 80% leverage ratio).

The U.S. earnings stripping rules do not prohibit the legal structure. And while the U.S. transfer pricing rules impose "arm's length" pricing requirements on intercompany transactions, IRS agents do no commonly focus on shareholder loans (in my experience). Unlike Sheppard, I don't fault section 163(j), so much as lax or unsophisticated U.S. tax administration.

German Rules (and Copyists)

Historically, the U.S. has exported principles of tax law and administration. But the rest of the world has caught up. Sheppard summarizes discussion at a recent International Bar Association annual meeting in Dubai. Tax professionals explored various European earnings stripping limitations. At some point, U.S. policymakers will start paying attention.

According to Sheppard, Germany was the first country to revise its earnings stripping rules. The German rules have serious teeth. The rules apply to all debt of a German affiliated group (an organschaft). The deduction for net interest expense is limited to 30 percent of EBITDA, regardless of whether the payer is related to the lender. EBITDA for this purpose is capped by the group's taxable income. (This effectively limits the benefit of timing or temporary differences to an organschaft with external or internal leverage.)

Practitioners apparently regard the limitation as "stingy." In my view, this is another example of German sensibility and efficiency. The rule is simple to describe and understand. It keys off EBITDA of a business; the financial metric that lenders and equity investors care about when financing an enterprise. The limitation may get low marks on "elegance," but it gets high marks from an administrative perspective. From a tax compliance perspective, policymakers should balance ease of administration against "technical perfection."

Italy has adopted a close variation of the German rule. The French legislature is considering a similar regime to address perceived abuse of French thin cap rules. The Dutch, Irish and Swedish are openly struggling with excessive leverage in domestic capital structures due to cross-border private equity transactions.

If you work in the international tax space or the domestic M&A space, Sheppard's article is definitely worth a quick read.

$95 Million Pay Day

Several interesting developments this week converged to prompt this post.

- On Monday, our favorite tax fashionista, Lee Sheppard, published an article discussing the relationship between income tax rates and executive compensation. Sheppard argues that Reagan-era cuts to marginal income tax rates have fueled the dramatic inflation in pre-tax compensation payments to C-suite executives. See "News Analysis: Should We Adopt a Millionaire's Surtax?" 133 Tax Notes 307 (Oct. 17, 2011).

- Sheppard mines data compiled by Catherine Mulbrandon from a November 2010 report by Jon Bakija of Williams College, Adam Cole of the Treasury Department and Bradley Heim of Indiana University. The report analyzes the occupations of the top 1% on American taxpayers between 1979 and 2005.

- Sheppard focuses, in particular, on the profiles of the top 0.1% of American taxpayers. Catherine Mulbrandon presents a terrific graphic comparison of the top 1% and the top 0.1% on her website, Visualizing Economics.

- Catherine Rampell, an economics reporter at the New York Times, also discussed the report on Monday. Rampell's analysis contains some good tabular summaries of the underlying data. She gently tweaks the OWS protesters by noting that the large majority of taxpayers in the top 1% are not Wall Street financiers. They are doctors and lawyers and C-suite executives.

- Over the weekend, Kinder Morgan signed a deal to acquire El Paso Corp in a $38 billion transaction. If the transaction is completed, El Paso's current CEO will be eligible for an exit package worth $95 million. Perhaps he can purchase the Greek island of Santorini after closing the deal.

- Today, Paul Caron posted the following chart summarizing the percentage of income earned and income taxed paid across the tax brackets in 2009. Thanks to Peter Pappas for linking. (Note that the data compiled for the Bakija et al study dates to 2005, so the numbers aren't precisely apples to apples.) The average taxpayer in the top 1% earned a comfortable $343,000 in 2009. That's approximately 0.4% of the retirement package for El Paso's current CEO. The doctors and lawyers in the top 1% are no closer to the top 0.1% than the middle-class union members in the middle of the income distribution.


How does this all fit together?

Sheppard argues that American CEOs are "grotesquely overpaid," and I completely agree. According to Sheppard:

Chief executive pay has increased 500 percent in real terms since the 1981 act cut the top marginal rate to 50 percent. Average worker pay has not increased in real terms during the same period. The median chief executive of an S&P 500 company now takes home more than $9 million annually [versus $1 million in 1981, in inflation-adjusted dollars]. The average chief executive of a public company takes home $2.5 million. Research has calculated that executive pay is the equivalent of 10 percent of corporate earnings at the largest companies.

It really is outrageous. I'm a strong believer in free market economics, but executive compensation practices are symptomatic of market failure. No manager of an industrial/energy company deserves a $95 million retirement package. Bill Gates, Mark Zuckerberg ... different stories.

Consultants in the executive compensation space make a killing on "peer benchmarking." For a public company, the board of directors is responsible for hiring and compensating the CEO. Directors hire consultants to "benchmark" the compensation packages granted to CEOs at "peer" companies with somewhat comparable operations, capital structures, etc. The "benchmarking" studies are used to support ever-increasing compensation packages to CEOs. As Sheppard observes, "[n]o board wants to believe that its chief executive is merely average."

The net result is the ridiculous inflation in executive compensation that has occurred during the last 30 years. The benchmarking data contains a "one-way" bias, because boards never pay their executives less than the "average" CEO among the comps. If nobody is ever "below average," there is never a price mechanism to slow the price inflation for executive management. As CEO compensation increases, the consultants incorporate the increased compensation data into their benchmarking analyses. Year after year, the cost of an "above average" manager continues to inflate. The newest vintage of "above average" CEOs needs to be paid more than the last vintage of "above average" of CEOs.

Ironically, CEO pay seems to be the one area where business organizations seek to maximize their costs. Imagine how long a business would last if it committed to paying "above average prices" for the various inputs into its products or services. Competitive markets reward innovative enterprise that can satisfy consumer demand while lowering costs of production. I understand that CEOs play a critical strategic role within their respective businesses. They're distinguishable from commoditized widgets. Nonetheless, corporate boards are consistently failing their shareholders by acquiescing to current executive compensation trends.

(Boards have other incentives to inflate CEO pay. Frequently, CEOs install or support directors who are themselves CEOs of other public companies. This creates an obvious conflict of interest. A CEO of Company A that wears a "director hat" for Company B is more likely to support an outrageous compensation package for the CEO of Company B, because he or she knows that the compensation package will become part of the benchmarking data supporting his or her compensation as CEO of Company A.)

Sheppard discusses various options to curb the inflation of executive compensation, including surtaxes and new brackets for the highest-paid taxpayers. I don't believe that taxes created the problem, so I don't believe that taxes will solve the problem. Instead, corporate boards should adopt guidelines on the relationship between CEO compensation and the median compensation of an employee within the organization. It could be a simple formula: CEO pay for a given period may not exceed 10x or 20x of the median employee pay. It could be more complicated, with performance metrics and clawbacks. No matter how you slice it, corporate boards should begin chipping away at the market failures that are inflating CEO pay.