The Ever-Updated Personal Saving Rate
The Bureau of Economic Analysis (BEA) estimates that the personal saving rate for the first quarter of 2013 was 2.3 percent - a five-year low, and a substantial drop from the fourth quarter of 2012, when it stood at 5.3 percent. Since many economists think a healthy household balance sheet is a necessary condition to fuel a stronger economic recovery, should we be worried about how low this estimate of the saving rate is?
We argue that the answer to this question is no, at least not yet. Quarterly saving rates are fairly volatile, and even though the first estimate for April came in at an equally paltry 2.5 percent, we should wait to see whether such low readings are confirmed in the next few months. More importantly, though, initial estimates for the personal saving rate normally end up being substantially revised. Moreover, these revisions are overwhelmingly on the positive side; that is, the final estimates are usually a lot higher than the initial ones. How much higher? The initial estimate for the personal saving rate has averaged 4.9 percent since World War II, while the final (current) estimate is 7 percent. So when we say revisions are substantial, we are not exaggerating.
Why is the personal saving rate so hard to estimate? The BEA computes the personal saving rate as part of its National Income and Product Accounts (NIPA) and defines it as the ratio of personal savings to disposable income. Personal savings, in turn, are obtained by subtracting personal outlays (consumption expenditures, interest payments, and current transfer payments) from disposable personal income, which is personal income minus personal current taxes.
This is where things get tricky. While the BEA has a very good handle on personal outlays, disposable income is considerably harder to define and estimate. Here are its main components:
- Compensation of employees (wages and salaries plus employer contributions to pension plans and social insurance)
- Proprietors' income (the income of owners of nonincorporated businesses)
- Rental income
- Income receipts on assets (interest and dividend income)
- Current transfer receipts (from Social Security, Medicare, etc., but also from businesses) net of contributions
While some of these components are straightforward to estimate, particularly the ones involving government outlays and receipts, others are inherently hard to define. Moreover, some income sources that have become fairly important for households in the last 30 years, like capital gains on equity and real-estate, are excluded altogether.
The revision process is typically a lengthy one. Data for a given quarter are first published in an advance release late in the first month of the following quarter. After that, the second and third (aka final) estimates are published one and two months after that, respectively. Then, usually in the following summer, the latest three years of data are revised, so that the estimates typically undergo three rounds of annual summer revisions. After that, estimates are only revised in benchmark revisions, when the BEA reconsiders its definitions and classifications to more accurately portray an ever-evolving economy, and it introduces new and improved statistical methodologies. Such benchmark revisions are usually very substantial and occur every four years. One is coming up in July this year.
There is an alternative way of obtaining estimates for the personal saving rate using the Flow of Funds Accounts (FOFA) reported by the Federal Reserve Board. It is based on the fact that savings (income minus outlays) are simply changes in net worth. The FOFA and NIPA concepts of savings actually differ in that the former includes net expenditures in consumer durables while the latter does not. Nonetheless, the FOFA also reports a NIPA-concept equivalent savings using FOFA data. The resulting saving rate is very noisy, so we show 8-quarter moving averages in the figure below. In contrast to the NIPA saving rate, the FOFA saving rate is not only higher, it has been increasing.
The lesson is that we should be careful when making inferences about household deleveraging based on the latest BEA estimates for the saving rate. Not only are these usually subject to substantial revision, but at this time alternative measures of the saving rate are pointing in a different direction.