Working Paper: NBER ID: w0968
Authors: Jacques Mairesse; Alan K. Siu
Abstract: Using a multivariate autoregressive framework, we have found a simple causal structure for the variables of interest q, s, r, and i, which is consistent with our data. As expected from the stock market efficiency hypothesis, q, the stock market one period holding rate of return, is exogenous relative to the other three variables (or Granger causes them). As postulated in the traditional accelerator model of investment, the rate of growth of sales, s, can be also treated as exogenous to the rates of growth of R&D and physical. investment, r and i. Moreover, no strong feed- back interaction is detected between the last two (r and i). Within the simple structure of the extended accelerator model, the substantive conclusion is that R&D and physical investment react very similarly to the growth of the sales and to movements in q; the response of R&D is, however, more stable or less irregular than that of physical investment. Expected demand and expected profitability thus both appear to be important determinants for R&D expenditures and physical investment.
Keywords: R&D investment; physical investment; multivariate autoregressive model; accelerator model; expected profitability
JEL Codes: O32; D24; E22
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| stock market one-period holding rate of return (q) (G17) | growth rates of sales (s) (C69) |
| stock market one-period holding rate of return (q) (G17) | growth rates of R&D (r) (O39) |
| stock market one-period holding rate of return (q) (G17) | physical investment (i) (E22) |
| growth rates of sales (s) (C69) | growth rates of R&D (r) (O39) |
| growth rates of sales (s) (C69) | physical investment (i) (E22) |
| expected demand and expected profitability (J23) | R&D expenditures (r) (O32) |
| expected demand and expected profitability (J23) | physical investment (i) (E22) |