This paper seeks to address the stock price adjustment toward fundamentals. Using the class of Switching Transition Error Correction Models (STECMs), we show that two regimes describe the dynamics of stock price deviations from fundamentals in the G7 countries over the period 1969-2005. Deviations appear to follow a quasi random walk in the central regime when prices are near fundamentals (i.e. transaction costs being greater than expected gains, the mean reversion mechanism is inactive), while they approach a white noise in the outer regimes (i.e. transaction costs being lower than expected gains, the mean reversion works). As expected when transaction costs are heterogeneous, the STECM shows that stock price adjustments are smooth, implying that the convergence speed is time-varying according to the size of the deviation. Finally, using appropriate indicators, both the magnitudes of under- and overvaluation of stock price and the speed of the mean reversion are exhibited per date in the G7 countries, showing that the dynamics of stock price adjustment is highly dependent on the date and on the country under consideration.