A recent National Bureau of Economic Research report on Section 936, the tax law that gave U.S. multinational corporations enormous tax breaks in the 20th century, has drawn some questionable conclusions.
“US Multinationals in Puerto Rico and the Repeal of Section 936 Tax Exemption for U.S. Corporations” by Zadia M. Feliciano and Andrew Green argues that the repeal of Section 936 decreased average manufacturing wages by 16.7%, and also reduced the number of manufacturers in Puerto Rico by 18.7% to 28.0%.
Section 936 was phased out from 1995 to 2006, after government analysts had concluded that the tax cuts offered corporations were costing the federal government more than they benefited Puerto Rico. It was also found in a number of studies that Section 396 had not increased the number of jobs available in Puerto Rico. Companies like Microsoft were able to attribute large percentages of their income to Puerto Rico while hiring small numbers of people on the Island. The report holds that the repeal of 936 was “driven by the desire to increase U.S. tax collections.” In fact, the Congress’s Joint Committee on Taxation (among others) specifically explained at the time that 936 was repealed because it had turned out not to be an effective means of encouraging job growth.
Using data from the Internal Revenue Service, (including the phase-out period), the authors determined that the amount of tax credits taken by companies under Section 936 began to decrease in 1983, before the phase-out began, and continued to do so up to 2005, before the phase-out ended. Between 1983 and 1995, when the phase-out began, the credits claimed fell from more than $4.5 billion to $3 billion. The continuing decline from 1995 to 2005 was a further $2 billion plus — in other words, the decrease after the phase out began was not strikingly different from the rate of decrease before the phase-out began.
Since the authors acknowledge that there is “no clear relationship between the value of tax credits received by U.S. corporations and their expenditures on salaries and wages” under 936, it is not clear that the reduction in credits claimed had any specific effect on the people of Puerto Rico. “It is difficult to establish a causality,” the authors lament at one point, referring to the changes in the amount and number of tax credits claimed.
The report also shows that manufacturing employment increased between 1983 and 1997 (after the phase-out) and then decreased from 1997 to 2012 (six years after the final phase-out). However, it is not just manufacturing employment that decreased during this time period, but all employment. Even if we presume that the drop in credits actually reflects the levels of employment in pharmaceuticals manufacturing, by far the most important vertical affected by 936, that does not explain the drop in employment in construction and other important industries in Puerto Rico.
The report is flawed in that it attributes all the decrease in manufacturing wages to the phase-out, ignoring free trade agreements, environmental and labor regulations, and the cost of electricity, among other factors. By choosing 1983 as a start date and a range of end dates, the authors have made their conclusions even less convincing.
They also used data from the U.S. Economic Census, comparing Puerto Rico with certain states and with the U.S. as a whole.
For Puerto Rico and the broader U.S., that report states, “The results show no significant impact of the elimination of Section 936 on the number of establishments, value added, and employment in Puerto Rico.” That is, when the U.S. is used as a control group, the changes in Puerto Rico do not show significant correlation with the phase-out of Section 936.
The selected control group states were Indiana, North Carolina, Oregon, and New Jersey. This group of states, like the U.S. as a whole, when used as a control group did not support the claim that the end of Section 936 caused significant changes in Puerto Rico.
When New Jersey was removed, however, the authors were able to find statistically different results between Puerto Rico and the three other states. The reasons for choosing the four states initially seem somewhat flimsy. Indiana and South Carolina have large amounts of jobs in manufacturing. However, Puerto Rico arguably showed a lot of manufacturing income but not a high proportion of jobs in manufacturing.
Oregon was chosen because it has a high percentage of employment in the area of food and beverage production, but no more than Texas and less than California. New Jersey was chosen because it’s big in pharmaceuticals — but was then rejected when it failed to show the desired statistical results. It might be interesting to choose other sets of states and see what the data shows.
Economist Ramon Cao has said the the natural economic system in Puerto Rico was “broken” once incentives for industrialization were given, tax reforms were incomplete, and changes to the dynamics of public spending resulted in the collapse of the Government’s assets. Looking at economic trends under those circumstances and attempting to correlate them with a single area of legislation may be impractical.
Economist Alejandro Silva argued that Puerto Rico needs to develop strategies to curb inequality, including the structure of large tax benefits to the corporate sector. The territory of Puerto Rico, he points out, has a lower effective corporate tax rate than all of the countries in the Organization for Economic Cooperation and Development.
Manuel Laboy Rivera, Secretary of the Puerto Rico Secretary of the Department of Economic Development and Commerce, says that as a matter of policy, the possible effects of the Section 936 phase-out may be a moot point.
“The economic development model has remained the same. Before and after section 936, Puerto Rico has relied on economic incentives to grow,” he said. “However, no economy can continue moving forward only through incentives.”
Some additional points to consider in evaluating the report:
- The estimated impacts are of a significantly lesser magnitude than those often cited in the territory;
- The number of plants is not the best indicator of the effect on the economy;
- The study did not report the changes in the value of production, which for example increased by about a third during the last decade of depression in the territory.
- Another major miss was not recognizing that as the amount (value) of 936 credits decreased companies were paying even less tax by operating their PR plans as foreign corporations.