Recoveries as a third phase of the business Cycle, based on the article by Antonio Fatás, Ilian Mihov 14 August 2013 The last recession in the US ended in June 2009. Yet, three years on, unemployment remains high. This column argues that we need to better understand how business cycles of recession and expansion work. Detailed evidence from the US suggests that recoveries are not simply mirror images of recessions. Because of its policy relevance, economists and policymakers must acknowledge that the pattern of recession/recovery has significantly changed over the last half century. According to the NBER business cycle dating committee, the last recession in the US ended in June 2009 (NBER 2013). Three years later US unemployment remains high and most estimates suggest that output remains below potential; a pattern also present in other advanced economies. As a result, central banks have made explicit commitments to keep interest rates at low levels until the recovery is firmly established, referring to a future date when the economy is close enough to ‘normal’. Despite the interest and policy relevance of measuring and understanding the years that follow a recession, we tend to describe business cycles as the succession of recessions and expansions, without separating the earlier years of an expansion – what we refer to as the recovery – and the later years. In a newly released CEPR discussion paper (DP 9551) we have produced an explicit analysis of the recovery phase for the US business cycles since 1950. Our approach is based on the original work of Burns and Mitchell that described a multi-phase business cycle that included a revival phase in between the recession and the expansion. We explicitly identify and date this third phase, what we will call the recovery. Identifying the recovery We work with a framework that borrows from the academic literature to combine the following elements: The growth path of an economy can be described by a long-term trend. Business cycles are seen as cyclical (transitory) deviations from this trend. We think of the trend as the maximum level of output. This is different from the interpretation of potential output as a ‘sustainable’ level of output in the tradition of measures such as the output gap or the non-accelerating inflation rate of unemployment (the NAIRU). Our notion of trend is similar to that of the Friedman’s plucking model (Friedman 1964). Business cycles are asymmetric and driven by negative shocks. This asymmetry is implicit in the Burns and Mitchell or the National Bureau of Economic Research methodology because of their focus on recessions.1 After a trough, there is a distinct phase where the economy returns towards normal levels (trend). This phase is called ‘recovery’. Defining and dating a recovery phase is a challenge because it requires a definition of what it means to return to ‘normal’ (defined as full employment or trend output)