Working Paper: NBER ID: w4888
Authors: Martin Feldstein; James H. Stock
Abstract: This paper examines the problem of measuring the growth of a monetary aggregate in the presence of innovations in financial markets and changes in the relationship between individual assets and output. We propose constructing a monetary aggregate so that it is a good leading indicator of nominal GDP; in general the weights on its components vary over time. We investigate two specific procedures: one in which subaggregates discretely switch in and out, and one in which the growth of the aggregate is a time-varying weighted average of the growth of the subaggregates, where the weights follow a random walk. These procedures are used to construct aggregates which potentially augment M2 with stock and/or bond mutual funds. Over 1960-1991, the time-varying aggregates look much like M2, but during 1992-93 the time-varying aggregates outperform M2.
Keywords: monetary aggregates; nominal GDP; financial innovations
JEL Codes: E41; E52
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Optimal control of M2 by the Federal Reserve (E52) | Reduction in long-term average rate of inflation (E31) |
| Optimal control of M2 by the Federal Reserve (E52) | Reduction in variance of nominal GDP growth (E19) |
| Newly defined monetary aggregate (E51) | Stronger and more stable leading relationship with nominal GDP (E19) |
| Redefining monetary aggregate (E19) | Improve explanatory power with nominal GDP (E10) |
| Rate of change of M2 (E49) | Rate of change of nominal GDP (E20) |
| Rate of change of M2 + Short-term interest rates (E43) | Rate of change of nominal GDP (E20) |
| Time-varying aggregates (C22) | Outperform M2 in predicting nominal GDP growth (E19) |