Cash For Clunkers Raises August SAAR From 9.5 to 16 Million And Other Observations
The first fully blown subsidy program in the form of Cash for Clunkers is now over. Many more undoubtedly will follow, with Cash for Woodshed Plasma TV likely to appear in 3 to 6 months. As it will facilitate Made in Taiwan and Japan exports to the US, one does not anticipate any WTO trade wars. Anyway, here is what Goldman had to say about the impacts of CfC:
1. Monthly vehicle sales: 15+ million SAAR in August. Of the vehicles sold under the “cash for clunkers” program, almost 600,000 were sold in August; this translates to about 7 million at an annual rate on a not-seasonally adjusted basis. Even after accounting for the fact that August typically has about 10% higher sales volume than the average month, this could have pushed the monthly seasonally adjusted annual rate (SAAR) of sales over 15 million vehicles, perhaps over 16 million. One question is whether customers who couldn’t take advantage of the rebate kept purchasing at a similar rate in August as in prior months (about 9 ½ - 10 million SAAR).
And after August's 16 million SAAR which is a 60% boost to July numbers, who wants to guess what September's SAAR will plunge to?
2. Non-auto sales: probably not much cannibalization. The intense focus on vehicle sales amid a generally difficult consumer environment may have come at the cost of spending in other areas. Although it will be impossible to know for sure even after the data are released, we suspect that non-auto spending will be slightly lower in August than it would otherwise have been. This observation is based on a very back-of-the-envelope analysis: we look at two other months in the last decade in which heavy promotions caused the monthly SAAR to jump by more than 2 million (October 2001 and July 2005). In both months, non-auto retail spending was slightly weaker than predicted (an average of 0.3 percentage points) by a model incorporating hours worked, weather variables, gasoline prices, and several other factors. If all spending was affected similarly, the impact would be 0.3 percentage points of nonauto consumer spending, or about $2½ billion.
Alas, in both 2001 and 2005 the Consumer saving rate was nowhere near as close to where it is currently. The possibility of non-auto cannibalization this time around is substantially higher.
3. Consumer spending: a 2-point boost to annualized growth in Q3. The nominal value of cars sold under the CARS program in July and August was probably about $15 billion (this assumes an average pre-rebate price per car of almost $22,000; the mix of vehicles purchased under the program appears to cost a bit less than the US average). Netting out perhaps $1-2 billion for cannibalization of other sales per point (2) above, this represents about 0.5% of quarterly consumer spending at current levels. In that case, the boost to the sequential growth rate of real consumer spending would be about 2 percentage points at an annual rate. (The implications for real and nominal spending diverge a bit, since consumers themselves are spending less—because of the rebate—but are still getting the same cars.) Our forecast for the third quarter is for 2% annualized growth in real consumer spending, so we’re essentially assuming that nonauto spending is flat.
The administration is desperate to front load GDP improvements at all costs: is the 25 pts increase in the S&P really worth it? Also begs the question what is the impact on car prices for the above mid-range and premium segments? Are BMW and Audi starting to really sweat? Don't look now, but dare we say it, deflation?
3. Inventories: less contribution to growth in Q3, more in Q4. Insofar as the boost to vehicle sales was more rapid than expected, manufacturers may end up with less inventory than they anticipated. The rate of inventory depletion will almost certainly be less than in Q2 – so there will still be an inventory-based contribution to GDP growth from the vehicle sector – but some of this contribution seems likely to be shifted into the fourth quarter.
Toyota which was the primary beneficiary of CfC announced yesterday it is reducing global capacity by 10%. Nuff said. Next.
4. Real GDP: Help in the second half. The cash for clunkers program will boost consumption and deplete inventories in Q3, relative to what would have been the case. These have opposite implications for Q3 GDP. If the boost in sales from the program was a complete surprise to automakers, then the consumption and inventory effects would be offsetting, with no impact on GDP until automakers increased production to rebuild inventory levels. In reality, automakers knew the program was coming, and had already scheduled substantial production increases in Q3. Most of the large manufacturers announced further production increases, mostly in Q4, after the early success of the program became evident. So the GDP effect will be split between the two quarters. Overall, if we take $15 billion of incremental spending as a summary of the impact, this would be worth about 0.2% of GDP in the second half of the year. If the effect were evenly split between the two quarters, it would raise the level of GDP 0.2 percentage points in each quarter; this would imply a 0.8% percentage-point boost to annualized sequential growth in Q3 and no impact in Q4 (since the level of activity would be the same).
Today's declining inventory number was conclusive evidence that the
inventory build story, and resultant GDP boost, is still a myth. Next.
5. Consumer prices: a small discount in August. The value of the rebates in August represents approximately 18% of the vehicles’ retail price. That discount applied to about 10% of the vehicles sold in July, but should apply to around 40% of the vehicles sold in August, for a net change of -5.4% ( =(40%-10%)*-18%) in the average vehicle price. New vehicles have a weight of about 4.5% in the CPI, so the net impact on the overall level of consumer prices should theoretically be 4.5% * -5.4%, or a decrease of 24 basis points. Note that this is only a temporary effect on the price level, so there should be a full payback for the July and August effects in September, when consumers will no longer benefit from the rebates. Also note that this could be partly or fully offset even within the month if dealers chose to offer lower discounts in light of the government rebate (this seems to have been what happened in July, when new vehicle prices in the CPI posted a substantial month-on-month increase despite the rebates).
A decline in CPI is just what the doctor ordered: one hopes Bernanke can spearhea,d an inflation campaign on increasing beer prices alone.