On the effects of policy uncertainty on stock prices

For a theoretical model on the effects of government policy on stock return, see Pastor and Veronesi (2012).

For more details on constructing this index, visit: www.​policyuncertaint​y.​com.

Others have also used the new policy uncertainty measure to assess its impact on other macro variables. The list includes Wang et al. (2014) who assessed the response of corporate investment to the new uncertainty measure, Pastor and Veronesi (2013), Ko and Lee (2015), and Brogaard and Detzel (2015) who assessed the response of risk premia and market returns to uncertainty, Baker et al. (2016) who considered the response of economic activity and firm-level outcomes, Bahmani-Oskooee and Ghodsi (2017) who investigated the response of housing prices in each state of the U.S., Kang and Ratti (2013) as well as Bahmani-Oskooee et al. (2018) who looked into the link between the new uncertainty measure and oil prices, and Bahmani-Oskooee et al. (2016) who assessed the impact of policy uncertainty on the demand for money in the U.S.

This could easily be observed if we solve (1) for ε_t and lag the solution by one period.

For the exact procedure of deriving normalized estimates see Bahmani-Oskooee and Tanku (2008).

It should be indicated that Pastor and Veronesi (2013) assessed impact of the same policy uncertainty measure on S & P 500 volatility and return using a general equilibrium model of government policy choices. Our short-run findings for the U.S. are consistent with them as well as with Brogaard and Detzel (2015).

It should be mentioned that no monthly GDP data were available. At the suggestion of a referee, we simply

use the same quarterly number to scale all of the monthly M2 data for each quarter. This is based on the assumption that quarterly GDP data do not change much month -to –month. Furthermore, for brevity we only report the short-run effects of policy uncertainty and not other variables. Complete results in five tables are available upon request from corresponding author.

Similar results are also reported by Ko and Lee (2015) who used Wavelet analysis and not error-correction modeling or cointegration.

The remaining statistics are similar to those of the multivariate models.