Long-run monetary neutrality

Abstract

In this article, we provide a test of long-run monetary neutrality employing cointegration and vector error-correction modelling methodology. Using quarterly data for the United States, we estimate the long-run relationships among money supply and output and other key macroeconomic variables. Our findings, in general, raise doubts about the long-run monetary neutrality proposition.

Notes

¹See for example, Mankiw (2000) and O'Sullivan and Scheffrin (2003).

² Since many of the series appear to have some kind of trend and none of the series have zero mean, we must allow for a trend in the data. Further, the unit root tests performed indicate that none of the series are trend stationary. Therefore, we do not restrict the trend to be in the cointegrating relationship. Thus, we allow for a deterministic trend in the data and include an intercept but no trend in the cointegrating relationship. Further, the main results of the article are the same regardless of the specification of the deterministic components in the VAR.

³ For a discussion of the procedure see Johansen (1991) and Johansen (1995).

⁴Boschen and Mills (1995) used 7 lags in the level data (6 lags in the differenced data). King and Watson (1997) also used 7 lags in the level data (6 lags in the differenced data), but reported some additional results for 5 and 9 lags in the level data (4 and 8 lags in the differenced data).