Written by Steven Hansen
February 2012 Real Personal Consumption Expenditure (PCE) – the inflation adjusted spending of consumers improved month-over-month, but Disposable Personal Income (DPI) again declined slightly:
- Real PCE (inflation and seasonally adjusted) improved 0.5% month-over-month
- Real DPI rose (inflation and seasonally adjusted) declined 0.1% month-over-month
- the personal savings rate (expressed as a percentage of DPI) fell 0.5% to 3.7%
- The market looks at current values (not real) expecting a PCE rise of 0.6 to 0.8% (versus 0.8% actual), and a rise in DPI of 0.3% (versus 0.2% actual).
Overall, this data seems mixed. On the bright side, consumers are returning to the trough. The bad news is that the inflation adjusted data is saying consumers do not have increased income to spend. Econintersect believes year-over-year trends are very revealing in understanding economic dynamics. Again, there was a broad, and in places significant, revision this month in the data for the last six months, which is explained below (see caveats below).
Keeping it real, per capita inflation adjusted income was $32,716 in February 2007 – and is down slightly at $32,600 in this February 2012 data. Roughly, expenditures are growing 1.5% faster than income since 2007.
The graph below illustrates the relationship between income (DPI) and expenditures (PCE). The consumer since mid 2010 has continued to spend more of his income (and therefore saving less). Simply, the consumer is continuing to return to the trough.
PCE is the spending of consumers. In the USA, the consumer is the economy. Likewise, personal income is the money consumers earn to spend. Even though most analysts concentrate on personal expenditures because GDP is based on spending, increases in personal income allow consumers the option to spend more.
There is a general correlation of PCE to GDP. This index has shown negative growth several times since the end of the 2007-09 recession. The overall trend since August 2009 is slightly negative. PCE is a fairly noisy index and subject at times to significant backward revision (see caveats below). Econintersect views this data using a three month moving average which is now becoming “less good”.
The summary table follows, and Econintersect uses the inflation adjusted (chained) numbers. Disposable Personal Income (DPI) is the income left after the tax man.
And please note that previous analysis of PCE and DPI has been somewhat negated by backward revision:
Estimates have been revised for October 2011 through January 2012. Changes in personal income, current-dollar and chained (2005) dollar DPI, and current-dollar and chained (2005) dollar PCE for December and January — revised and as published in last month’s release — are shown below.
Personal savings rate jumped significantly in December, but now has declined again in January and February 2012. In an economy driven by consumers, a higher savings rate does not bode well for increased GDP. This is one reason GDP may not be a good single metric of economic activity. The question remains what is the optimal savings rate for the current demographics. It might be expected that as people near retirement, the savings rate rises and after people retire, savings rate falls. Econintersect is not aware of any study which documents this effect. The graph below is from BEA table 2.6.
Caveats on the Use of Personal Income and Consumption Expenditure Data
PCE is a fairly noisy index and subject at times to significant backward revision. This index cannot be relied upon in real time.
This personal income and personal consumption expenditure data by itself is not a good tool to warn of an upcoming recession. Econintersect has shown that PCE is a distraction for recession watchers, with moves over a few months having a 30% accuracy of indicating a recession start, and a 70% incidence of indicating a non-recessionary event. The graph below shows the lack of correlation. Note, however, that PCE does have prolonged declines over many months associated with recessions but these long declines are not very good in “predicting” a recession until it is already underway.
Readers are warned that this article is based on seasonally adjusted data. Monthly non-adjusted data is not available. Econintersect has concerns that seasonally adjusted data is not accurate in the New Normal.
The above graph plots year-over-year data instead of month-over-month which is likely a more accurate approach to understanding PCE. Again, this is seasonally adjusted data – and there should NOT be continuing occurrence of year-over-year single month blips every few months (red circles on above graph).