Fiscal Policy

Warning: long, wonky, and not for the fainthearted. I try to assess how the international reforms will impact where firms book profits and thus the measured trade and income balance, not just the mechanical impact of a higher fiscal deficit.

The eurozone member-states may never be able to run a fiscal policy that is optimal for the eurozone taken as a whole.

Abe's consumption tax hike stalled what until then had been a demand led recovery, leading to a three year period with no growth in domestic demand. 

This post is authored by Peter Erickson, a staff scientist at the Stockholm Environment Institute and a co-author of a new paper in Nature Energy that studies how much of U.S. oil reserves are economical to extract as a result of government subsidies that benefit the oil industry. This post is a response to a previous post, by Varun Sivaram, arguing that federal tax breaks for the oil industry do not, in fact, cause a globally significant increase in greenhouse gas emissions, citing a recent CFR paper authored by Dr. Gilbert Metcalf at Tufts University. Dr. Sivaram’s short response to Mr. Erickson’s rebuttal is included at the bottom of the post. As Congress moves towards tax reform, there is one industry that hasn’t yet come up: oil. While subsidies for renewable energy are often in the cross-hairs of tax discussions, the billions in federal tax subsidies for the oil industry rarely are; indeed, some subsidies are nearing their 100th birthday. And yet, removing oil subsidies would be good not only for taxpayers, but for the climate as well. The lack of attention on petroleum subsidies is not for lack of analysis. Congress’ own Joint Committee on Taxation values the subsidies at more than $2 billion annually.  (Other researchers have put the total much higher.) Just in the last year, two major studies have assessed in detail how these subsidies affect investment returns in the US oil industry. The two analyses—one published by the Council on Foreign Relations (CFR) and the other in Nature Energy (which I coauthored)—both show the majority of subsidy value goes directly to profits, not to new investment.   That inefficiency—both studies argue—is reason enough for Congress to end the subsidies to the oil industry. But oil subsidies also have another strike against them: oil is a major contributor to climate change. The burning of gasoline, diesel, and other petroleum products is responsible for one-third of global CO2 emissions. That climate impact is one of the reasons the Obama Administration had committed, with other nations in the G7, to end these subsidies by 2025. Both the CFR and Nature Energy analyses arrive at a similar figure as to the net climate impact. As CFR fellow Varun Sivaram notes in a previous post on this blog comparing the two studies, the CFR study finds that subsidy removal would reduce global oil consumption by about half a percent. Our analysis for the Nature Energy study also finds a reduction in global oil consumption of about half a percent. (You won’t find this result in our paper, but it is what our oil market model, described in the online Supplementary Information, implies.) The most critical place where the studies—or rather, authors—differ is how they put this amount of oil in context. (Our study also addresses many more subsidies, and in much more detail, than the CFR study, but that is not the point I wish to address here.) Sivaram refers to the half-percent decrease in global oil consumption as “measly…washed out by the ordinary volatility of oil prices and resulting changes in consumption…the nearly-undetectable change in global oil consumption means that the climate effects of U.S. tax breaks are negligible.” I would argue that this assertion confuses the effect of subsidies on oil consumption with our ability to measure the change. But before I get into it further, let me first describe how much oil and CO2 we are talking about. By the CFR paper’s estimates, removal of US oil subsidies would lead to a drop in global oil consumption of 300,000 to 500,000 barrels per day (corresponding to 0.3% to 0.5% of the global oil market). The sequential effects in their model are shown in the chart below, which I made based on their results. It shows their lower-end case, in which global oil consumption drops by 300,000 barrels per day (bpd). (This case is described in their paper as using EIA’s reference case oil price forecast and an upward-sloping OPEC supply curve.)  In their model, a drop of over 600,000 bpd in US supply from subsidy removal is partially replaced by other sources of U.S., OPEC, and other rest-of-world supply, yielding a net reduction in global consumption of roughly half as much (300,000 bpd, shown in the right column). (This ratio itself is also interesting and important. For each barrel of oil not developed because of subsidies, this case shows a drop in global oil consumption of 0.45 barrels. The CFR study’s other three cases show a drop of 0.51, 0.63, and 0.82 barrels of global consumption for each US barrel left undeveloped.) Each barrel of oil yields, conservatively, about 400 kg of CO2 once burned, per IPCC figures. So, the range of impacts on oil consumption in the CFR study (again, reductions of 300-500k bpd or 110 million to 200 million bbl annually) implies a drop in global CO2 emissions of about 40-70 million tons of CO annually. (The actual emissions decrease from subsidy removal could well be greater, because this estimate doesn’t count other gases released in the course of extracting a barrel of oil, such as methane or other CO2 from energy used on-site). From a policy perspective, 40 to 70 million tons of CO2 is not a trivial (measly) amount. Rather, it is comparable in scale to other U.S. government efforts to reduce greenhouse gas emissions. For example, President Obama’s Climate Action Plan contained a host of high-profile measures that, individually, would have reduced annual (domestic) greenhouse gas emissions by 5 million tons (limits on methane from oil and gas extraction on federal land), 60 million tons (efficiency standards for big trucks), and 200 million tons (efficiency standards for cars). The CFR authors don’t quantify their findings in CO2 terms, however, and Sivaram refers to oil market volatility as a way to discount CFR’s findings on reduced oil consumption, concluding that the effects are “undetectable” and “negligible.” The argument is essentially that because other changes in the oil market are bigger, and can mask the independent effect of subsidy removal, that subsidy removal has no effect on climate change. This line of argument conflates causality, scale and likelihood of impact (which in this case are either all known, or can be estimated) with ability to monitor, detect and attribute changes (which is rarely possible in any case, even for more traditional policies focused on oil consumption). By this logic, almost any climate policy could also be discounted as immaterial, because it is rare to be able to directly observe with confidence both the intended result of a policy and the counterfactual – what would have happened otherwise.  Rather, I would argue that if we are to meet the challenge of global climate change, we’ll need these 40 to 70 million tons of avoided CO2, and many more, even if there is uncertainty about exactly how big the impact will be. Concluding an action represents a small fraction of the climate problem is less a statement about that action than it is about the massive scale of the climate challenge. Indeed, as the Obama White House Council on Environmental Quality stated, such a comparison is “not an appropriate method for characterizing the potential impacts associated with a proposed action… because…[it] does not reveal anything beyond the nature of the climate change challenge itself.” So, I argue that subsidy removal is indeed material for the climate, even by the CFR report’s own math. And as Sivaram also notes, the CO2 emission reductions would multiply as other countries also phase out their subsidies. Lastly, I need to disagree with Sivaram’s statement that our study is “written in a misleading way”. He asserts this because in the Nature Energy article we focus on the entire CO2 emissions from each barrel, rather than apply an oil market economic model as described above that counts only the net, or incremental, global CO2. But the incremental analysis method above is not the only way to describe CO2 emissions. Indeed, comparing the possible CO2 emissions from a particular source to the global remaining carbon budget is a simple and established way to gauge magnitudes, and nicely complements the incremental analysis enabled by oil market models.   As another noted subsidy expert—Ron Steenblik of the OECD—commented separately in Nature Energy, our analytical approach provides an important advance because it enables “researchers to look at the combined effect of many individual subsidies flowing to specific projects and to use project-specific data to gauge eligibility and uptake.” Similar assessments of other countries, and other fossil fuels, would provide an important window on the distortionary impacts of these subsidies and their perverse impacts on global efforts to contain climate change. Sivaram Response to Erickson Rebuttal First of all, I am grateful to Peter Erickson for responding in this way to a blog post I wrote that was critical of his conclusions. His response was graceful and sophisticated—I think I largely agree with it, and he’s pointed out some holes in my post that I want to acknowledge. However, I do still stand by my headline, “No, Tax Breaks for U.S. Oil and Gas Companies Probably Don’t Materially Affect Climate Change.” In fact, I think the Erickson rebuttal above reinforces just that point. Tackling the overall thesis first: in his rebuttal, Erickson is willing to accept that a reasonable estimate for the carbon impact of U.S. tax breaks for oil and gas companies is 40–70 million tons of carbon dioxide emissions annually (there may be other greenhouse gas emissions, such as methane, that increase the climate impact). Erickson even compares the magnitude of this negative climate impact with the positive impact of President Obama’s efficiency standards for big trucks. I am absolutely willing to accept that removing U.S. tax breaks for oil companies would be about as big a deal, in terms of direct emissions reduction, as setting domestic efficiency standards for big trucks. Importantly, this direct impact is trivial on a global scale, which is the point that I made in my original post, reinforcing Dr. Metcalf’s conclusion in his CFR paper. I am, however, sympathetic to Erickson’s argument that the world needs a rollback of tax breaks, efficiency standards for big trucks, and a whole suite of other policies in the United States and other major economies to combat climate change. And there is certainly symbolic value to the United States rolling back its oil industry tax breaks, possibly making it easier to persuade other countries to follow suit. I also want to concede that Erickson very rightly called me out on unclearly discussing the relationship between oil price volatility and the effect on oil prices of removing tax breaks. We definitely know which direction removing subsidies would move prices (up) and global consumption (down). I should have been clearer that my comparison of the frequent swings in oil prices to the tiny price impact of removing subsidies was merely to provide a sense of magnitude, NOT to imply that measurement error washes out our ability to forecast the magnitude of tax reform’s price impact, ceteris paribus. Finally, Erickson took issue to my characterization of his paper as “misleading.” Indeed, I never meant to imply that he and his co-authors intended to mislead anybody. I still, however, stand by what I meant: that the paper might lead a casual reader to take away an erroneous conclusion by relegating the global oil market model to an appendix and only citing the increase in U.S. emissions in the main body. In my opinion, readers need to know that industry tax breaks have a very small effect on global greenhouse gas emissions, but there are other very important reasons to remove them. And yes, the United States absolutely should remove its tax breaks, as should other countries remove their fossil fuel subsidies. On that count, Erickson and I are in complete agreement.

Goods made in China flow through Ireland before returning to China, and the many other ways tax influences the balance of payments.

Trade in tax often trumps trade in tasks.

I have a new paper out with David Kamin of New York University Law School—it will be formally out in Tax Notes in a couple of weeks, but given that there is a live debate on the topic, we are posting it in draft form now—and the New York Times had a related editorial linking to it earlier this week. So this is a joint post with David Kamin. Our paper makes two arguments. 1) Even with fairly optimistic assumptions about long-term growth and long-term interest rates and the persistence of “excess returns” from U.S. direct investment abroad, the U.S. cannot sustain trade deficits of approaching 3 percent of GDP over the long-run. The CBO’s estimates for long-run growth and the long-run nominal interest rate on the U.S. debt stock imply a sustainable long-run trade deficit of about 1 percent of GDP. That would generate maybe 20 basis points of GDP in permanent revenue. If the excess returns (“dark matter”) on U.S. foreign direct investment go away, the U.S. would need to run a small trade surplus—and the border adjustment would lose revenue over the long-term. As a result, realistic projections of revenue from a border adjustment should show that revenue falling considerably and, possibly, entirely disappearing over the long-term. Remember the border adjustment acts as a tax on imports (imports are not deductible as a cost) and a subsidy for exports (a portion domestic wage and other content of exports is effectively rebated, as exports are not considered revenues while domestic wages are considered expenses, creating a tax loss). So it only generates revenue in net if the revenue collected on the border adjustment on imports exceeds the revenue lost on the export rebate. This is actually standard old school IMF external debt sustainability analysis.* The basic logic is simple. Sustained trade deficits over time lead to a build-up of external debt, and it is impossible to borrow forever both to run trade deficits—as over time you have to borrow both to finance the trade deficit and the interest on your accumulated debt. Countries can run modest current account deficits over time without rising external debt, but they cannot run trade deficits over time (well, technically, the non-interest current account, not the trade deficit) without an ever-rising stock of external debt to GDP. For those more familiar with fiscal sustainability, the trade deficit is roughly analogous to the primary fiscal balance. A modest headline fiscal deficit is often consistent with a stable debt to GDP ratio, but a large primary fiscal deficit generally is not. An obvious irony: a border adjustment is the only source of the long-term revenue Paul Ryan needs to make the tax reform permanent through the budget reconciliation process if the U.S. is projected to run permanent trade deficits. Or in the terms of Trump’s campaign, if the U.S. keeps losing at trade—an overall deficit means lots of bilateral deficits. 2) The actual revenues that the border adjustment will generate will be a function of how transfer price and similar tax games evolve, not just how the trade balance evolves—and current estimates fail to take into account the fact that the border adjustment will be avoided to some significant degree. To be sure, it might very well reduce the considerable tax gaming that we see in the current system—but not by 100 percent. Exempting exports from corporate income tax (and providing a rebate against domestic costs) while taxing imports removes the incentives for some tax games, but creates incentives for others. Especially if the U.S. alone adopts a destination-based cash flow tax. The border adjustment would remove the incentive for some existing tax games. That is a attractive feature of the proposal. Some firms try to shift profits made in the U.S. outside the U.S. through creative transfer pricing—say importing drugs invented in the U.S. from Ireland for example, and booking the profit on U.S. sales in Ireland. A destination based tax would help there. It also would end firms’ needs to game the U.S. system when it came to export of intellectual property—probably the largest distortion in the current system. Notably, though, it would do so by giving up U.S. taxes of the global export of U.S. intellectual property and possibly turning the United States into a tax shelter from other countries’ source-based tax systems. Right now, tech companies do not so much avoid tax on their U.S. profits as avoid tax on their global profits—Apple Ireland gets the rights to Apples IP for Europe, Asia, the Middle East, and Africa, and those profits are effectively untaxed (at least until the funds are repatriated back to the U.S). A destination-based cash flow tax would entirely exempt revenues from the exports of the intellectual property from tax, so it wouldn’t matter if the profit was booked in Ireland or California (and if Apple’s domestic cost exceeded its domestic revenues, it might even get a rebate). Only the destination would matter from the U.S. perspective. We consequently would expect that exports of intellectual property would rise sharply with the tax reform, bringing down the reported trade deficit—in part because companies will try to avoid remaining source-based taxation of other countries. After all, if Germany still taxes based on source of the profit and the United States does not, it makes all the sense in the world to report that profit in the United States! However, this new gaming would almost entirely affect the revenue bases of other countries, not the revenue base of the United States. But the incentive to export—and for businesses to avoid importing—would create new kinds of tax games that do affect the U.S. tax base and that revenue estimates we’ve seen don’t seem to take into account. Specifically, there would be—as David Hemel has noted—strong incentives to book U.S. sales as exports where possible. A firm would export to an offshore subsidiary and deduct its exports from revenues. And then sell directly from abroad to consumers, avoiding any business tax. More generally, firms that import would try to sell directly to consumers if the tax was entirely collected through businesses. In order for it to work, you actually have to have “VAT” style border enforcement to avoid leakage. Here is an example, one Brad got from his father, who brought back a German car as a souvenir while on sabbatical in the 1960s: Go on a vacation in Europe. Pick up a BMW in Munich. Spend a couple of weeks touring. Put your new car on a ship—and fly back to the U.S. Then pick up your (almost) new car at the nearest port. You pay the shipping, but not the 20 percent border-adjustment. Not if the border adjustment is imposed through business tax filings. Think that won’t happen? Think again. They way to stop it is to do actual adjustments for any goods at the border. But then it isn’t just a business income tax. And policing virtual borders is even harder. Realistically some leakage is inevitable—and needs to be factored into the revenue estimates. Simply taking the current trade deficit and multiplying it by 20 percent (or whatever the new tax rate is)—what some current estimates do—is almost sure to over-estimate actual revenues. And then there’s the fact that trade deficits should dissipate over the long-term. * New school IMF debt sustainability analysis focuses almost entirely on fiscal debt sustainability. Brad thinks that is a step backwards.

Neil Irwin’s column on the border-adjustment tax spurred an interesting debate. Irwin notes that a 25 percent rise in the dollar (or even a somewhat smaller rise) would have an impact outside the United States, as the dollar is a global currency. Dean Baker and Jared Bernstein note that the dollar moves around a lot without destroying the global economy. The projected moves in the dollar are no larger than the dollar’s 20 percent or so move over the last three years. Baker: "Movements of this size happen all the time. They certainly can cause problems, but the financial system generally deals with it." Fair enough.* I still worry though. There is a difference between a 20 percent move (off a long-term low) and a 30 or 40 percent move. And there is a difference between normal exchange rate volatility and large, sustained currency shifts. I would note that the 20 percent rise in the dollar in late 2014 and early 2015 contributed to the change in China’s currency regime—and China’s shift to managing against a basket roiled global markets from August 2015 to February 2016. The world survived, but it wasn’t totally smooth. Let me focus on two specific reasons for concern: One. Balance sheet mismatches in emerging economies.** Two. Dollar pegs, or basket pegs heavily weighted to the dollar. One. Balance sheet mismatches. These are the old bane of emerging economies. Countries over-borrow in foreign currency. Their currency depreciates. And the country cannot pay its debts back. This was a big part of the story of the Asian crisis. And of subsequent defaults in Ecuador, Russia, and Argentina. But things have changed since the 1990s. The governments of most large emerging economies have more foreign currency assets than foreign currency debts. Central bank foreign exchange reserves exceed foreign currency bonds issued by the government. This is the case in China’s central government—it has $3 trillion in foreign currency reserves and almost no foreign currency denominated bonds. And the governments of Brazil, Russia and India all also have more foreign currency reserves than foreign currency debt. The Saudis have issued a lot of dollar debt recently. But the Saudi government is still net long the dollar—foreign currency bonds are $17.5 billion (plus some loans), reserves are still $536 billion. The governments of big emerging markets are generally not big borrowers in dollars anymore. Even those with current account deficits. There are some (partial) exceptions. Turkey is an interesting case, though the government of Turkey has only a modest amount of foreign currency debt (roughly $30 billion). But it is close because Turkey really has very few reserves. Around $55 billion of the central banks $106 billion in reserves (foreign currency and gold) have in effect been borrowed from the banks (see Paul McNamara; or the footnotes in Turkey’s SDDS reserves disclosure), and thus aren’t a hedge for the government. Argentina too after its recent bond issue. Yet, in broad terms, emerging market governments are better positioned to withstand big foreign currency moves than in the past. (On the flip side, foreign investors have taken on more local currency risk—they are the ones who stand to lose.) As many have noted, the FX risks within emerging markets now largely lie on the corporate side. Emerging market firms stepped up their dollar borrowing even as emerging market governments stepped down. I wouldn’t assume that just because firms were able to withstand a twenty percent move in the dollar they can easily withstand a forty percent move—especially if U.S. rates move up. The actual risk though likely varies. A lot of the biggest emerging market corporate issuers of dollar denominated bonds likely have state backing—PEMEX ($100 billion in debt, not all in dollars), Petrobras (over $100 billion in debt), Rosneft, Gazprom, the Chinese state banks, the big Chinese airlines, and the like. And most of these countries hold enough reserves to cover the foreign currency debts of their big state companies as well as the direct foreign currency debts of the state. But that doesn’t mean that a further rise in the debt burden of state companies wouldn’t be painful. Those most exposed are countries where there are big, broader foreign currency mismatches throughout the corporate sector. Turkey for example. Any rise in the dollar effectively shifts wealth from Turkey’s firms to Turkey’s households—at least those with dollar deposits.*** At some point the strain will become too big, and the banks will have to restructure their corporate loan book and absorb losses. Turkey’s banks are already hurting from a decline in the lira, which reduces their equity buffers (the banks have big foreign currency balance sheets, but keep their equity in lira—so the equity base shrinks relative to their total liabilities as the currency weakens). And there are a set of countries and companies that are exposed not so much to the dollar’s value relative to other large exporters of manufactures (Japan, China, Korea and Europe) but to the dollar price of commodities. That though is a subject for another time. The impact of the border adjustment on global commodity prices (notably oil) is a hotly debated topic, and I am not convinced that it will be as mechanical as some have argued. The other risk comes from the world’s remaining dollar pegs. A stronger dollar pulls up the value of a lot of other currencies, as many countries still peg to the dollar -- even though a reasonable number of de facto dollar pegs have either broken or transformed into basket pegs over the last two and a half years (the Kazakh tenge for example, or the Nigerian naira) I will set aside the questions around the Saudi peg. A stronger real effective riyal would hinder the Saudis’ efforts to diversify their economy away from oil, but I never took those efforts all that seriously. What really matters for Saudi Arabia in my view is not so much the value of the dollar against the G-3, but the value of oil in Saudi riyals. Dollar/oil matters more than dollar/yen—or put differently, dollar/yen only matters to the extent it changes dollar/oil. The hardest questions involve China. It doesn’t float now. And strictly speaking it doesn’t manage its currency against the dollar. Rather it tightly manages its currency against a basket. China would have to decide how it wants to react to a border adjustment. It won’t happen automatically. And it is easy to see how this could be an important source of future instability. Consider the following. China could decide to maintain a basket peg, and thus let the renminbi adjust to the border adjustment in proportion to the global market adjustment (and specifically moves in the dollar/euro and dollar/Asia). Because of the weight of the dollar, the Hong Kong dollar and the Saudi riyal in China’s basket, the renminbi would move by say 2/3rds as much as the dollar moves against the major advanced economies (Mexico and China dominate the dollar basket versus emerging markets—and neither enters into China’s own basket—China for obvious reasons and Mexico because direct trade is limited). So if the dollar appreciated by about 15 percent against the majors, the mechanical operation of the basket would produce a roughly 10 percent appreciation in the renminbi against the dollar. That would offset less than half the border adjustment on China’s exports (a full offset of a 20 percent border-adjustment on imports, given the math, requires a 25 percent move). And other countries would also gain relative to China, as their currencies would depreciate more. Would the market—which in China, is primarily domestic institutions that can move funds across the border more easily than international investors—believe that China wouldn’t move by more? Or would it start to speculate that China ultimately would want a full adjustment? Would China’s controls be effective if domestic exporters and importers believed that a further depreciation was imminent? And if China did a full adjustment—a one off reset of the level of the basket—how would other currencies respond? Would China trigger even bigger depreciations elsewhere in Asia? It doesn’t seem to me all that hard to see how there is an overshoot in Asia, given the large role that China plays in Asian trade. I know that goes against the conventional wisdom that the border adjustment won’t be perfectly offset because capital flows exceed trade flows, which is true, but incomplete—as it doesn’t quite capture the dynamics around America’s largest single source of imported goods and the biggest single contributor to the dollar’s trade-weighted index. China’s currency is heavily influenced by trade flows, including financial outflows disguised as trade flows—and of course by expectations about the yuan’s future course against the dollar. Those would be destabilized by a border adjustment. China doesn’t have giant domestic foreign currency mismatches. There are a few (the airlines) but they are manageable in the big scheme of things. The risk to China comes from the potential impact of outflows (and the policies introduced to stem outflows) on domestic financial stability. And the risk to the world in turns comes from the trade impact from a complete break in China’s peg and a major depreciation of the renminbi. One that overwhelms any border adjustment. * Small aside on the border adjustment: a border adjustment eliminates the incentive to game the system by shifting profits offshore, but it does so by exempting profits on exports from any onshore tax. It basically abandons the notion of trying to tax the intellectual property (IP) rents on export income, or the economic rents from the export of natural resources for that matter. And it could create incentives for other kinds of gaming. I am convinced that the current system is heavily gamed in ways that hurt U.S. exports of IP and high-margin products (the active ingredient in pharmaceuticals for example), but the proposed border-adjustment gets rid of some of the games by in effect not taxing certain kinds of hard-to-tax income. ** I am assuming that large financial institutions in the advanced economies have matched books (whether directly, or through cross-currency swaps and similar hedges), and thus do not have any worrying balance sheet mismatches. That is also just an assumption. At the same time, the moves in the dollar/euro and dollar/yen have put balance sheets to test. *** Turkey’s central bank publishes data on the foreign currency book of its corporate sector. Others should too.