Working Paper: NBER ID: w2294
Authors: Patric H. Hendershott; Joe Peek
Abstract: The official personal and private saving statistics contain a number of conceptual measurement errors. In this paper we develop and analyze personal and private saving measures adjusted for the difference between income tax payments and actual liabilities, saving via net purchases of government pension assets (including social security) and consumer durables, and that part of after-tax interest income attributable to inflation. We find that the adjusted personal and private saving rates in recent years are only slightly below their post-1950 averages, not at all time lows as reported in the official NIPA statistics. Furthermore, over the past 35 years, personal saving has been more volatile and corporate saving less volatile than the official measures. Also, the inflation premium corrections remove the negative correlation between personal and corporate saving. That is, the often observed negative correlation between the official measures of personal and corporate saving is due solely to measurement errors in the two series. Finally, the decrease in federal government saving in the 1980s is the continuation of a 30-year trend, not a one-time aberration.
Keywords: Private Saving; Measurement Errors; Personal Saving; Corporate Saving
JEL Codes: D14; E21
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| measurement errors (C20) | incorrect conclusions about saving behavior (D14) |
| measurement errors (C20) | significant understatements of personal saving (D14) |
| measurement errors (C20) | observed negative correlation between personal and corporate saving (D14) |
| inflation premium corrections (E31) | remove the negative correlation between personal and corporate saving (D14) |
| decrease in federal government saving in the 1980s (H62) | part of a longer-term trend (E32) |