Yesterday, when we presented the Bloomberg interview of Princeton economist and former Fed vice chairman Alan Blinder, we speculated that his statement that "more easing is necessary" was the first shot across the QE3 bow. Today, Goldman's Sven Jari Stehn has fired the second one in a paper just released titled: "Fiscal Adjustment without Fed Easing: A Tall Order" in which he basically takes our conclusion from the Blinder interview to the next level. As Blinder said previously, in order to improve the once again deteriorating labor picture, more fiscal stimulus would be necessary. That, however, is impossible, especially in a Congress where everyone is now promising $4 trillion of deficit cuts over the next few years. The only difference is how this cutting will be achieved: republicans want spending cuts, while democrats are demanding tax hikes for the richest. While neither approach will work in the US without the shock of a bond-crash induced austerity, Goldman conducts an thought experiment in which it evaluates the effectiveness of a tax-based and a spending-based fiscal consolidation. While finding that on average spending based deficit reduction is more effective, it only truly works in parallel with assistance from monetary policy: be it an interest rate decrease (impossible due to ZIRP) or further Large Scale Asset Purchase (QE) program. In other words, the only thing that can prevent an economic contraction in the next 2 years of semi-austerity, will be more monetary easing.
Furthermore, Goldman also openly admits that in either fiscal case, the drag on economic growth will be substantial. "A number of studies have shown that adjustments focused primarily on spending cuts (“spending-based consolidations”) tend to be notably more successful at delivering such large consolidations than revenue-based ones. Building on work done by the IMF, we identify two reasons for this difference. First, spending-based consolidations are usually more persistent, as they are often combined with structural reforms. Second, spending cuts tend to be less damaging for growth than tax increases...A key factor behind this difference in success, however, is the response of monetary policy. While spending-based adjustments are typically accompanied by monetary easing, tax-based ones often see monetary tightening. Using a counterfactual experiment which “shuts down” the interest rate response, we show that the difference in growth damage between spending and tax-based adjustments narrows sharply..With the funds rate close to zero, our analysis implies that both spending and tax-based consolidations are likely to act as a significant drag on growth. Nonetheless, spending-based adjustments might still be the lesser of two evils, particularly if combined with entitlement reform and fiscal rules that come with a strong enforcement mechanism." Translation: the economy will slow materially regardless, but without monetary easing it will crash. Next up: cue an enjoinder by the New Jersey installment of the Ivy League, and the balance of Wall Street, all of whom realize that their bonuses are suddenly at steak.
We said yesterday that "we believe that with this the opening salvo for more cash demands, which will be met with staunch opposition in D.C., thereby kicking the ball back to the Fed (which already is doing everything in its power to deflate all commodities as rapidly as possible - a trend which will sooner or alter engulf risk assets as well) the only alternative is monetary. Aka more quantitative easing. And when that becomes apparent, and when Goldman's Jan Hatzius is firmly on board, the full court press for another round of easing can begin." Well, Goldman just got on board. Look for the cries for more monetary intervention courtesy of a Congress which can't make up its mind about a debt ceiling hike for 4 months to escalate over the next 2-3 months as the economic reality turns aggressively south. At that point the Chairman will be faced with a daily barrage of "experts" who are screaming "deflation... or printing." We have a guess which one Ben will chose.
From Sven Jari Stehn of Goldman Sachs
I. Fiscal Adjustment without Fed Easing: A Tall Order
The US needs a very large fiscal consolidation as we expect a primary (ex-interest) deficit of 7.7% of GDP this year. Stabilizing the debt stock eventually requires the primary budget to be close to balance. Although some of this deficit is cyclical, the structural deficit (defined here as the primary deficit adjusted for the cycle and one-off accounting changes) currently stands at 6% of GDP. Moreover, this is the very minimum adjustment needed as stabilizing the debt stock at current levels—or even returning the debt ratio to pre-crisis ratios—would require a much larger fiscal consolidation.
Ingredients for a Successful Consolidation
A number of studies—going back to Alberto Alesina and Roberto Perotti in 1995—have identified factors that determine the “success” of a fiscal consolidation, which is typically defined as a sizable and lasting reduction of the deficit or debt ratio.
The key result of these studies is that spending-based consolidations tend to be much more successful than revenue-based ones. There are two suggested reasons for this result. First, these studies argue that spendingbased consolidations are usually more persistent because they are often accompanied by structural reforms. Also, spending-based adjustments tend to be politically more difficult and thus signal stronger commitment to continued fiscal consolidation than tax increases. Second, these studies find that spendingbased adjustments are less detrimental to growth—and indeed can boost growth. The better growth outcome eases the consolidation burden both directly (through higher tax revenues) and indirectly (because it makes it easier to sustain the adjustment).
Recent work by the IMF, however, suggests that these conclusions should be re-examined. First, the IMF has shown that all consolidations—whether spending or revenue-based—tend to act as a drag on growth when we look at intended consolidations (or consolidation efforts) directly instead of the resulting changes to the structural deficit. In particular, the IMF argues that existing studies “stack the deck” against finding significant adverse growth effects. By using the cyclically-adjusted budget deficit to identify fiscal consolidations, the earlier studies include episodes that were not genuine periods of adjustment but rather one-off accounting changes. Moreover, even when such one-offs are removed, the change in the structural budget deficit is often a poor proxy for deliberate changes in fiscal policy because it fails to detect attempted fiscal adjustments that result in sharp downturns and are therefore reversed quickly. Second, the IMF study suggests that monetary policy plays an important role in shaping the consequences of fiscal adjustment. Specifically, spending-based adjustments have a less detrimental growth effect than tax-based adjustments because they are, on average, accompanied by monetary easing while tax-based adjustments usually see monetary tightening. This suggests that the success of a consolidation in reducing the deficit or debt ratio might depend importantly on the monetary policy response.
Finally, the new IMF dataset allows us to explore to what extent intended adjustments actually result in expost improvements in the fiscal situation. That is, the dataset enables us to take into account that a consolidation attempt might have been so badly designed or implemented that
Spending Adjustments Are More Successful...
In an initial look at the IMF data, we organize the dataset into consolidation episodes. Specifically, we define a consolidation period as one in which policymakers intend to consolidate by at least 1% of GDP in the first and last periods. This definition produces 29 episodes of consolidation.
We then split these episodes into the five consolidations that produced the largest and smallest improvements in the primary balance. Exhibit 1 shows how effort and success vary across those two groups. Exhibit 2 lists details of these episodes. The two exhibits offer a number of interesting insights.
First, the required US adjustment is comparable only to the largest three consolidations in the dataset. Only Denmark starting in 1984, Sweden in 1993 and Ireland in 1982 have achieved consolidations in excess of 9% of GDP. Moreover, the most successful efforts were much more persistent than the unsuccessful ones (6 years on average versus 1 year).
Second, there is notable slippage between the adjustment effort and the actual improvement in the primary balance. During the least successful consolidations, the average adjustment effort (an average 1.5% of GDP) resulted in no improvement in the primary balance (Exhibit 1). But even for successful consolidations, the average adjustment effort (13.1% of GDP) was quite a bit larger than the reduction in the primary deficit (8.9% of GDP). One source of slippage is that the consolidation effort is only partially passed through into improvements in the structural deficit—most likely because some consolidation efforts were aborted before the fiscal outlook actually improved (e.g. Japan in 1997).
Another reason for slippage is the effect of the consolidation effort on growth, as discussed below.
Second, the exhibits confirm earlier studies that successful consolidations rely much more on spending reductions (around 72% of the adjustment) than unsuccessful ones (37%).
Finally, we see that the most successful consolidations, on average, saw no decline in growth while the least successful ones experienced a sharp 2.7 percentage point (pt) drop. At the same time, however, the top adjustments were accompanied by notably more monetary easing than the failures. During successful consolidations policy rates fell, on average, by 5.5pts while they only declined by 1.1pts during the least successful consolidations.
… Mostly Due to Monetary Policy
Given this very different response of monetary policy, we now explore the extent to which this drives the differences in success. To do so, we estimate a statistical model for the 15 countries between 1980 and 2009 that explains the joint dynamics of the IMF’s measure of intended consolidations, real GDP, the primary balance and the monetary policy rate. Once estimated, we use the model to trace out the effects of an intended consolidation on growth, the primary balance and the policy rate. To distinguish between the effects of spending and tax-based consolidations we estimate two additional models and plot the results alongside the average consolidation.
Exhibit 3 shows how the primary balance, on average, responds to an intended fiscal consolidation. Consistent with the evidence above, our estimates imply that a 1% of GDP consolidation effort typically improves the primary budget by only half that amount. Moreover, Exhibit 3 confirms that spending-based adjustments tend to be notably more successful in raising the budget surplus than tax-based ones. The reason is twofold. First, spending-based consolidation efforts tend to be sustained for longer than tax-based ones. Specifically, a 1% of GDP spending adjustment effort is usually followed by an additional spending cut effort of 0.4% in the year after, while a comparable revenue adjustment is only followed up with another 0.1% of GDP tax raise. As a result, spending-based adjustments raise the structural balance by almost twice as much after five years than tax-based adjustments (not shown).
Second, spending-based adjustments are less damaging for growth (Exhibit 4). Consistent with the IMF study we find that a 1% of GDP consolidation effort, on average, reduces real GDP by around ½% after two years. The composition of the adjustment, however, matters crucially: the GDP hit is much larger for tax-based consolidations (around 1½%) than spending-based ones (¼%).
A notable difference between spending and tax-based adjustments, however, is the response of monetary policy (Exhibit 5). The former are accompanied by monetary easing, while tax-based adjustments typically see monetary tightening in the first year (followed by some easing in the second year). The IMF shows that the initial monetary tightening is driven by interest rate hikes in response to indirect tax increases. A likely explanation is that central banks are worried about second-round inflation effects from increases in indirect taxes.
Consolidating Without Monetary Policy
Taken at face value, these results suggest that spending cuts are an overwhelmingly more attractive option than tax increases: they tend to be more persistent and less damaging to growth (although they don’t raise growth as suggested by some previous studies). Applying this conclusion to the required US adjustment, however, would be naïve because the above results likely overstate the success that can be expected from spending-based adjustments relative to tax-based ones in the current environment. With the funds rate close to the zero lower bound, a spendingbased adjustment could not be accompanied by monetary easing unless the Fed decided to adopt another asset purchase program (which we think is highly unlikely). Moreover, the Fed would most likely see through any indirect tax increases—were they to occur—and probably not raise interest rates in response to a revenue-based consolidation.
In a counterfactual analysis we therefore attempt to “shut down” the interest rate response to get a better sense of the implications of the choices the US currently faces. Such an experiment is fraught with difficulty as it requires an estimate of how changes in the policy rate affect output and the budget deficit. To obtain such an estimate we proceed in two steps. First, we use quarterly data to estimate the effect of shocks to monetary policy on growth and the primary pioneered by David and Christina Romer. Second, we transform these estimates into annual data and apply them to the cross-country results above to construct the “no monetary policy” scenario. Given these steps, the uncertainty surrounding the resulting simulation is substantial. Moreover, allowing no monetary policy response is an extreme assumption as Fed officials could adopt additional unconventional policy steps to support a spending-based adjustment if they chose to do so.
With this in mind, Exhibits 6 and 7 suggest that the difference in success between spending and tax-based adjustments is less pronounced when there is no monetary policy response. In particular, the hit to GDP is now more similar during the first two years, at 1-1½% for tax and spending-based adjustments. After the second year, however, spending-based adjustments are still quite a bit less damaging to growth. As a result, the achieved improvement in the primary balance is now also more similar across spending and tax-based adjustments during the first two years. A 1% of GDP consolidation effort now improves the primary deficit ratio by half that amount after two years regardless of the composition. Spending adjustments, however, remain more effective at reducing the deficit persistently.
Spending Cuts are No Panacea, but Necessary
Our analysis implies that even spending-based adjustments are likely to be challenging in the current environment. This is because without a monetary response, both spending and tax-based consolidations are likely to act as a sizable drag on growth.
That said, our analysis suggests that spending-based adjustments are nonetheless likely to be the lesser of two evils. First, the difference in growth damage between spending and tax-based consolidations narrows sharply without monetary policy but remains noticeable three or more years after the start of the adjustment. Second, spending adjustments tend to lead to more persistent deficit reductions than tax increases. This is probably because such consolidations often involve entitlement reform and not just one-off reductions in discretionary outlays. More broadly, this highlights the desirability of multiyear fiscal rules that come with a strong enforcement mechanism.
Meanwhile, barring another round of asset purchases, the best the Fed can do is keep monetary policy on hold to cushion the growth drag from the fiscal consolidation—independently of whether it comes through adjustments in spending, revenue or both. As a result, the looming fiscal adjustment should reasonably be expected to see policy rates—and probably longer-term rates too—at lower than normal levels for an extended period.