Hedging Their Bets

On Wednesday Goldman Sachs, after picking a peculiar time to do so, decided to lower their Q3 GDP estimates, thereby lowering the bar for broad GDP expectations, and the resultant "beat" by the official BEA data of a 3.5% reading sparked the biggest market rally since July (only to be followed by an even bigger drop on Friday). Whether or not Goldman's prop trading operation benefiting directly or indirectly from this increase in volatility is unknown as the firm does not provide that level of granularity and detail in its earnings reports. Yet based on the most recent disclosure from the NYSE, Goldman is now trading at a more than 10:1 ratio of principal to agency (read gambling with taxpayer funds about 10 times more than it transacts on behalf of clients).

Zero Hedge is still hoping that the NYSE will eventually disclose what percentage of Goldman's principal trades goes to the exchange's SLP program, as previously promised. Perhaps one of the reasons why the NYSE's profitability is collapsing is because of the ongoing posture of opacity, despite all claims by Mr. Niederauer to the contrary. That, and of course dark pools encroaching more and more into exchange territory until such time as only retail sheep investors are left to trade on open venues.

Regardless of whatever P&L may have been generated by GS over the last two days when volatility took a roughly 35% round trip, Messrs. McKelvey and Hatzius are now hedging their bets on their GDP call, providing an extended matrix as to why not only will they be proven right, but why Q4 GDP will be a dramatic disappointment, courtesy of a CfC-esque pull forward of future GDP production/consumption.

As Rosenberg and many other strategists have been repeatedly banging on the table on this issue, the Q3 GDP is simply a recreation of the Cash For Clunkers effect, taken to the national level. Just as CfC managed to pump the SAAR rate for August to 14.3 million, which subsequently plummeted to 9 million in the next month, expect a comparable pattern with Q4 GDP. And unless the administration plans on essentially running the economy on one-time stimuli until the next presidential election, the bigger the artificial sugar high, the greater the subsequent crash. As much as Summer, Bernanke et al wish they could change the laws of economics (and physics), at this point there is no place to hide.

Goldman agrees with this simplistic assessment:

"How much of the rebound in real GDP was due to the fiscal stimulus, and where do we stand in terms of the effects of stimulus thus far?  Although precise answers are impossible at this juncture, several aspects of the report are consistent with our estimates that the fiscal package enacted in mid-February as the American Recovery and Reinvestment Act (ARRA) would have accounted for virtually all of the growth reported for the third quarter."

As Zero Hedge claimed, President Obama is widely encouraged to take a remedial analyst class at none other than Goldman Sachs. Perhaps, finally, he will understand that it is not Edmund's fault, nor are eveil economists to blame, when they point out what the significiance of one-time, non-recurring events is. And in this particular case, the cost of the 3.5% GDP artificial high reading is roughly $800 billion in incremental debt that is not going away, and that will have to be serviced: initially at a blended rate of about 2.75%, and potentially rising to 10% (just ask Paul Volcker).

Full GDP observations from Goldman Sachs:




Fiscal Stimulus Helps Third-Quarter Growth…

Gauging the precise effect of fiscal stimulus on third-quarter growth is difficult for two reasons.  First, the ARRA law had many provisions—in our analysis we split it into more than 50 line items—whose effects are scattered throughout the national income accounts.  Second and more importantly, we cannot know what would have happened in the absence of the stimulus.
That said, we can see clear patterns in the third-quarter data that, in conjunction with prior developments, suggest a strong helping hand from Uncle Sam:

1. The “cash for clunkers” program, which was not actually part of ARRA, spurred a large increase in purchases of new motor vehicles.  Real outlays on new vehicles rose to $191 billion (bn) at an annual rate in the third quarter from $154bn in the second quarter, as shown in Exhibit 1.  This accounted for slightly more than two-fifths (1.1 percentage point) of 2.5% annualized increase in real final sales.  Although we cannot know for sure how much of the increase in vehicle sales was due to cash for clunkers, the monthly data strongly suggest that most of it was.  The program was in effect from July 24 to August 24; consistent with that timing, both unit sales and the motor vehicles and parts component of real spending spiked in August after a more modest increase in July, and the September data show pull-backs to levels consistent with pre-clunker behavior.

Fiscal Stimulus: Passing the Peak Effect on Growth

Yesterday the Bureau of Economic Analysis (BEA) marked the onset of recovery in the US economy as it published its provisional estimate that real GDP rose 3.5% at an annual rate in the third quarter.  This was in the upper half of the range of private forecasts and well above our 2.7% estimate.  The report featured rebounds both in consumer spending (+3.4% in real terms) and in residential investment (+23.4%).  Real business investment and state and local spending fell, but by less than we had expected (-2.5% and -1.1%, respectively).

How much of the rebound in real GDP was due to the fiscal stimulus, and where do we stand in terms of the effects of stimulus thus far?  Although precise answers are impossible at this juncture, several aspects of the report are consistent with our estimates that the fiscal package enacted in mid-February as the American Recovery and Reinvestment Act (ARRA) would have accounted for virtually all of the growth reported for the third quarter.  Unfortunately, those same estimates also suggest that the growth impact of fiscal stimulus has passed its peak, absent significant extensions or new initiatives.  Without these, the fiscal contributions to real GDP growth will subside between now and mid-2010, after which we expect them to become drags.  Growth at that point will then require at least a modest pace of job creation to replace the support that Uncle Sam is now providing.


The effect on real GDP is more problematic, as some of this demand was undoubtedly met by drawing down inventories or by imports.  Indeed, real imports of autos and trucks surged more than $27bn at an annual rate last quarter.  However, domestic vehicle output also rose sharply, accounting for just about half (1.7 percentage points) of the increase in real GDP.  Although some of this increase in output would have occurred in any event, it is hard to avoid the conclusion that the cash for clunkers program was an important contributor.

2. Consumer spending on other items probably got a lift last quarter from tax cuts and extensions of benefit programs implemented during the first half of 2009.   While investment and infrastructure projects got much of the attention when ARRA was passed last winter, a more immediate purpose of this act was to shore up income-strapped state and local governments and households—in the latter case via tax cuts and enhancements to unemployment insurance and income support programs.  The effects on personal disposable income are clear; as shown in Exhibit 2, since February increases in government transfer payments and reductions in taxes have largely offset the effects of large and persistent declines in labor compensation and in income on assets on disposable income.  While the government support did not rise further in the third quarter, the level of this support remained high and essentially provided the wherewithal for last quarter’s increase in spending, as other components of income continued to erode on balance.  Given these patterns and the impact of the cash for clunkers program, it is not too much of a stretch to attribute most of last quarter’s increase in real consumer spending to fiscal stimulus.



3. The homeowner tax credit helped spark an upturn in residential investment.  This provision of ARRA provides an $8,000 tax credit—an effective price cut of nearly 5% on the median-priced existing home—to first-time homeowners who buy homes by November 30.  It has been widely credited for helping lift home sales from the historical depths to which they fell last winter.  Although some rebound was apt to occur from these lows, the timing and magnitude of this year’s gains in home sales and housing starts—increases of more than 20% since January—probably owe at least something to this tax credit.

If so, then last quarter’s increase in real GDP would show the effect in two places.  First, brokerage commissions are part of residential investment.  The “other structures” component that includes them rose $7.3bn at an annual rate in the third quarter, or $6.6bn in nominal terms.  This latter figure compares closely to the $7.1bn increase that would be implied from the sales data themselves.  Neither one is purely due to the tax credit—“other structures” include manufactured homes and major home improvements as well, while some of the increase in existing home sales might not have been due to the tax credit—but it surely played a role.  And if it did, then it would also be responsible for at least part of the $14.8bn annualized increase in single-family homebuilding that more than accounts for the rest of the rise in real residential investment, as multifamily construction continued to sag.

4. State and local spending would likely have fallen more if not for ARRA.  As noted above, states and localities were also specifically targeted by ARRA in an effort to soften the tax increases and spending cuts that these jurisdictions had to make to balance their operating budgets.  As with the personal income accounts, a large part of this support shows up as a sharp increase in federal grants-in-aid beginning in the first quarter of 2009, as shown in Exhibit 3.   Without this help, it is unlikely that real state and local spending would have rebounded as it did in the second quarter or dipped as little as it did this past quarter, which marked the first quarter of fiscal 2010 for most jurisdictions.  Notably, this increase was concentrated in construction outlays, which rose almost $13bn in the second quarter and nearly $2bn further in the third.



5. Business fixed investment may have gotten a lift from the expiring depreciation bonus.   Given the depths to which capacity utilization has fallen and the difficulties many firms have experienced in obtaining funds, the 2.5% annualized decline reported for real business investment in the third quarter is surprisingly mild.  One possible reason is that some companies accelerated the purchase of equipment in anticipation of the expiration of the depreciation bonus, which was extended to the end of 2009 in ARRA.

Taken as a whole, these observations imply that the US economy would have continued to contract last quarter in the absence of fiscal stimulus.  Note, for example, that gains in consumer spending and housing activity added 4.2 percentage points to last quarter’s growth; in other words, without them real GDP would have contracted by another 0.7% (the same as in the second quarter).  With fiscal stimulus providing most of the impetus to these increases. plus other components of business and state and local spending, it is hard to see how real GDP could have increased without these programs.

…But the Effects Diminish From Here

Unfortunately, our estimates also suggest that the peak of the fiscal support to US growth is now behind us, at least as the law now stands.  As shown in Exhibit 4, we estimate that the effects of ARRA will remain positive but subside through the second quarter of 2010.  Thereafter, this program exerts a drag on growth as the provisions that are now helping expire and pull GDP down by more than others lift it.


This statement often elicits surprise, for what we think are two reasons.  First, when the latest round of fiscal stimulus was first discussed about a year ago, the administration’s focus on proposals for infrastructure and investment projects received disproportionate attention in the press and in financial markets relative to what was actually enacted; in turn, this created a presumption that the effects of fiscal stimulus would build slowly given the long lead times for such projects.  Second, the effect of fiscal stimulus on GDP growth depends on the changes in the level of GDP the stimulus induces, and not on the level itself; this confusion is exacerbated by a focus on how little of the stimulus has yet been “spent” by the government.

To illustrate the first point, Exhibit 4 subdivides our estimates of the growth contribution from ARRA into three items—(1) income support to individuals (e.g., extension of unemployment benefits) plus aid to states and localities, (2) tax cuts to individuals and businesses, and (3) investment projects.   The first two categories comprised about two-thirds of the total package according to the final scoring of ARRA, and they took effect more quickly, accounting for seven-eighths of the fiscal 2009 total.  As a result, they have been responsible for the lion’s share of the growth effect thus far.

The second point is illustrated in Exhibit 5, where we plot the estimated effect of the entire bill on both the level of GDP (the line) and its growth rate (the bars).  The two are linked by the fact that the growth effects depend on how quickly the effect on spending is rising (the slope of the line).  Thus, the growth effects will tend to occur more quickly than would be suggested by statements to the effect that only a small portion of the money has been spent.  For example, while only 25% of the money has been spent according to recovery.gov, we have already seen more than half of the positive effect on real GDP growth, which should ultimately boost the level of GDP by about $370bn, or roughly 2½%, by mid-2010.  Thereafter, the growth effect turns negative as some of the initial provisions run out and the level of GDP declines as a result.



Of course, this is subject to change as Congress and the administration consider whether to extend various provisions as their expiration dates approach.  For example, as the homebuyer tax credit is due to expire at the end of November and several other provisions (unemployment benefits, health insurance for those who are unemployed, and various business tax breaks) come due in December, a “mini stimulus bill” is quickly coming together for consideration next week in the Congress.  So far, the effects look small (additional stimulus of about 0.3% of GDP in fiscal 2010) and temporary (no meaningful effect after fiscal 2010).  However, we cannot rule out more such efforts in coming months, at least until the labor market shows more signs of improvement than it has to date.

Ed McKelvey