Working Paper: NBER ID: w3612
Authors: Peter Rappoport; Eugene N. White
Abstract: Standard tests find that no bubbles are present in the stock price data for the last one hundred years. In contrast., historical accounts, focusing on briefer periods, point to the stock market of 1928-1929 as a classic example of a bubble. While previous studies have restricted their attention to the joint behavior of stock prices and dividends over the course of a century, this paper uses the behavior of the premia demanded on loans collateralized by the purchase of stocks to evaluate the claim that the boom and crash of 1929 represented a bubble. We develop a model that permits us to extract an estimate of the path of the bubble and its probability of bursting in any period and demonstrate that the premium behaves as would be expected in the presence of a bubble in stock prices. We also find that our estimate of the bubble's path has explanatory power when added to the standard cointegrating regressions of stock prices and dividends, in spite of the fact that our stock price and dividend series are cointegrated.
Keywords: stock market; bubble; premia; 1929 crash
JEL Codes: G12; N22
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| spread of interest rates (E43) | stock prices (G12) |
| bubble's path (E32) | standard cointegrating regressions of stock prices and dividends (C29) |
| shifts in preferences for savings and risk (D15) | stock prices (G12) |
| anticipated changes in economic policy (E69) | stock prices (G12) |
| premium on brokers' loans (G21) | stock prices (G12) |
| bubble's dynamics (E32) | stock prices (G12) |