3rd quarter GDP Estimate: Personal Income, Outlays, Construction

August Personal Income up 0.2%; 2 Months PCE Would Subtract 0.07 Percentage Points from Q3 GDP, Blogger and Commenter RJS reports at MarketWatch 666

The August report Personal Income and Outlays from the Bureau of Economic Analysis gives us nearly half the data that will go into 3rd quarter GDP, since it gives us 2 months of data on our personal consumption expenditures (PCE), which accounts for nearly 70% of GDP, and the PCE price index, the inflation gauge the Fed targets, and which is used to adjust that personal spending data for inflation to give us the relative change in the output of goods and services that our spending indicated . . . this report also gives us monthly personal income data, disposable personal income, which is income after taxes, and our monthly savings rate . . . however, because this report feeds in to GDP and other national accounts data, the change reported for each of those metrics is not the current monthly change; rather, they’re seasonally adjusted amounts expressed at an annual rate, ie. they tell us how much national income, spending, and savings would change over a year if August’s adjusted income and spending were extrapolated over an entire year…however, the percentage changes are computed monthly, from one month’s annualized figure to the next, and in this case of this month’s report they give us the percentage change in each annualized metric from July to August . . .

Thus, when the opening line of the news release for this report tells us “Personal income increased $35.5 billion (0.2 percent) in August,” they mean that the annualized figure for seasonally adjusted personal income in August, $20,716.7 billion, was $35.5 billion, or almost 0.2% more than the annualized personal income figure of $20,681.2 billion for July.

The actual, unadjusted change in personal income from July to August, which would be on the order of one-twelfth of that size, is not given . . . similarly, annualized disposable personal income, which is income after taxes, rose by 0.1%, from an annual rate of $ 18,105.0 billion in July to an annual rate of $18,123.9 billion in August . . . the monthly contributors to the increase in personal income, which can be seen in the Full Release & Tables (PDF) for this release, are thus also annualized . . . in August, the reasons for the $35.5 billion annual rate of increase in personal income were a annualized $47.1 billion increase in wages and salaries, and a $21.7 billion increase in government social benefits to individuals, which were partly offset by an annualized $28.5 billion decrease in business and farm proprietors’ income..

For personal consumption expenditures (PCE), BEA reports that they increased at a $130.5 billion annual rate, or by more than 0.8 percent, as the annual rate of PCE rose from $15,791.7 billion in July to $15,922.2 in August; that was after the July PCE figure was revised from the originally reported $15,832.3 billion billion annually, while prior months were revised as well, which were already included in the concurrent 3rd estimate of 2nd quarter GDP…..total personal outlays for August, which includes interest payments and personal transfer payments in addition to PCE, rose by an annualized $152.9 billion to $16,413.0 billion annually, which left total personal savings, which is disposable personal income less total outlays, at a $1,710.9 billion annual rate in August, down from the revised $1,824.6 billion annualized personal savings in July . . . hence, the personal saving rate, which is personal savings as a percentage of disposable personal income, fell to 9.4% in August, down from 10.1% in July..

As you know, before personal consumption expenditures can be used in the GDP computation, they must first be adjusted for inflation to give us the real change in consumption, and hence the real change in goods and services that were produced for that consumption . . . the BEA does that by computing a price index for personal consumption expenditures, which is a chained price index based on 2012 prices = 100, and which is included in Table 9 in the pdf for this report….that index rose from 115.849 in July to 116.314 in August, a month over month inflation rate that’s statistically 0.4014%, which BEA reports as a 0.4% increase, following the rounded +0.4% change in the PCE price index they reported for July…applying that August inflation adjustment to the nominal 0.8% change in August spending left real PCE up a rounded 0.4% in August, after a real PCE decrease of 0.5% in July . . . note that when those price indexes are applied to a given month’s annualized PCE in current dollars, it yields that month’s annualized real PCE in chained 2012 dollars, which are the means that the BEA uses to compare one month’s or one quarter’s real goods and services produced to another . . . the BEA’s result for that is shown in table 7 of the PDF, where we see that August’s chained dollar personal consumption total works out to 13,691.0 billion annually, 0.42323% more than July’s 13,633.3 billion, a difference that the BEA rounds down and reports as +0.4%…

To estimate the impact of the change in real PCE on the change in GDP, month over month changes do not help us much, since GDP is reported quarterly . . . thus we have to compare July and August’s real PCE to the the real PCE of the 3 months of the second quarter . . . while this report reports real PCE for each of those months separately, the BEA also provides the annualized chained dollar PCE for those three months quarterly in table 8 in the pdf for this report, where we find that the annualized real PCE for the 2nd quarter was represented by 13,665.6 billion in chained 2012 dollars…(note that’s also what’s shown in table 3 of the pdf for the revised 2nd quarter GDP report) . . . then, by averaging the annualized chained 2012 dollar figures for July and August, 13,633.3 billion and 13,691.0 billion respectively, we get an equivalent annualized PCE for the two months of the 3rd quarter that we have data for so far….when we compare that average of 13662.15 billion to the 2nd quarter real PCE of 13,665.6 billion, we find that 3rd quarter real PCE has shrunk at a 0.10% annual rate for the two months of the 3rd quarter that we have…{note the math we’ve used to get that annual contraction rate:(((13,633.3 + 13,691.0) / 2) / 13,665.6) ^ 4 = 0.99899, which we subsequently subtract from 1 }…that’s a pace that would subtract 0.07 percentage points from the growth rate of the 3rd quarter, should there be no improvement in September’s real PCE from that July & August average…

Construction Spending Unchanged in August after both June and July Spending were Revised 1% Higher

The August report on construction spending (pdf) from the Census Bureau estimated that our seasonally adjusted construction spending construction spending for the month was at an annual rate of $1,584.1 billion, which was statistically unchanged (±1.0 percent)* from the revised annualized estimate of $1,584.0 billion in construction spending in July, but was 8.9 percent (±1.5 percent) above the estimated annualized level of construction spending of August of last year . . . for the first eight months of this year, construction spending amounted to $1,034.5 billion, which was 7.0 percent (±1.0 percent) above the $966.7 billion spent over the same period in 2020 . . .

A further breakdown of the different subsets of construction spending is provided in a Census summary, which precedes the detailed spreadsheets:

• Private Construction – Spending on private construction was at a seasonally adjusted annual rate of $1,242.2 billion, 0.1 percent (±0.5 percent)* below the revised July estimate of $1,243.7 billion. Residential construction was at a seasonally adjusted annual rate of $786.6 billion in August, 0.4 percent (±1.3 percent)* above the revised July estimate of $783.5 billion. Nonresidential construction was at a seasonally adjusted annual rate of $455.6 billion in August, 1.0 percent (±0.5 percent) below the revised July estimate of $460.2 billion.
• Public Construction – In August, the estimated seasonally adjusted annual rate of public construction spending was $341.9 billion, 0.5 percent (±1.6 percent)* above the revised July estimate of $340.3 billion. Educational construction was at a seasonally adjusted annual rate of $79.8 billion, 1.1 percent (±2.0 percent)* above the revised July estimate of $78.9 billion. Highway construction was at a seasonally adjusted annual rate of $98.3 billion, 1.6 percent (±4.4 percent)* above the revised July estimate of $96.8 billion.

The construction spending report is used as source data for 3 subcomponents of GDP; investment in private non-residential structures, investment in residential structures, and government investment outlays, for both state and local and Federal governments . . . however, gauging the impact of revised July and August construction spending as reported here on 3rd quarter GDP is difficult because all figures given in this report are in nominal dollars and as you know, data used to compute the change in GDP must be adjusted for changes in price . . . accurately adjusting construction for price changes is not easy either, because the National Income and Product Accounts Handbook, Chapter 6 (pdf), lists a multitude of privately published deflators that are used by the BEA for the various components of non-residential investment, such as the Engineering News Record construction cost index for utilities’ construction spending….so in lieu of trying to find and adjust for all of those obscure price indices, we’ve opted to just use the producer price index for final demand construction as an inexact shortcut to make the price adjustment needed in order to make a ballpark estimate.

Producer price index shows aggregate construction costs rose 0.2% in August after rising 1.5% in July, 0.7% in June and by 0.5% from April to May . . . on that basis, we can estimate that construction costs for August were roughly 1.7% more than June, roughly 2.4% more than those of May and roughly 2.9% more than those of April, while obviously 0.2% more than those of July . . . we then use those percentages to inflate lower priced spending figures for each of the 2nd quarter months, which is arithmetically the same as adjusting higher priced July and August construction spending downward, for comparison purposes. . . . annualized construction spending in millions of dollars for the second quarter months is shown at $1,579,265 for June, $1,564,153 for May, and $1,553,547 for April in this report, while it was at $1,584,038 million for July and $1,584,085 million for August…thus to compare July and August’s inflation adjusted construction spending to that of the second quarter, our formula becomes: ((1,584,085 + 1,584,038 * 1.002) / 2 ) / (( 1,579,265 * 1.017 + 1,564,153 * 1.024 + 1,553,547 *1.029) / 3) = 0.98971, meaning real construction over July and August was 1.029% lower than that of the 2nd quarter….hence, that means that after adjusting for inflation,  real construction for the 3rd quarter fell at a 4.0538% annual rate from that of the 2nd quarter . . . that’s a contraction at a $16.489 billion annual rate, which means that if September should show no improvement, the pullback in real construction would subtract a net of about 0.44 percentage points from 3rd quarter GDP across those components that it influences…