通貨膨脹與產出、失業之間的關係一向是大眾所關心的問題，本文旨在檢定Mil ton Friedman在其諾貝爾經濟學獎演說中所提出之假說：通貨膨賬波動性會導致失業率上升及產出減少。 檢定Friedman假說所涉及的關鍵問題是如何衡量通貨膨脹波動性。本文根據美國、英國及西德1962年第1季至1988年第1季的資料，以“一般化自我迴歸條件變異數不齊一性” (GARCH)模型推估通貨膨賬波動性，並據以檢定Friedman假說。 本文之實證結果支持Friedman假說。由於通貨膨脹波動有損經濟制，經濟政策宜以降低此一波動性為導向。

This paper is motivated by Milton Friedman's hypothesis, proposed in his Nobel Lecture, that inflation volatility raises the rate of unemployment and lowers the level of output. Together with the casual empirical evidence that high inflation volatility is often accompanied by high inflation rates, the Friedman hypothesis amounts to claiming that the Phillips Curve is positively sloped. The crucial issue in testing the Friedman hypothesis is measuring inflation volatility. In this paper, I use a generalized autoregressive condi-tional heteroskedascity (GARCH) model to estimate the conditional mean and variance of inflation forecasts from time series data from the first quar-ter of 1962 through the first quarter of 1988, for the United States, Great Britain, and West Germany. The estimated conditional mean and variance are then employed to test the Friedman hypothesis. My empirical results imply acceptance of the Friedman hypothesis. To the extent that inflation uncertainty reduces economic welfare, policies designed to reduce that uncertainty are called for.