In this paper, we search for evidence of self-organized criticality (SOC) in the Brazilian aggregate production. In order to do so, long-term growth cycles are extracted from gross domestic product (GDP) data, by using the Singular Spectrum Analysis. Then, this times series is utilized in ARFIMA (p,d,q) regressions. Data set consists of quarterly observations from 1947 to 2012. As our results show, long-term growth cycles may range between 3 and 12 years, with an average duration of 9 years. Slowdown phases usually last 5 years, with a relatively low growth-rate of 2.9% p.a. By the other hand, expansions average 375 years, but have a higher rate of variation (6.9% p.a.). Additionally, volatility in Brazil has presented an inverted U-shape pattern, reaching a peak in the middle of the 1980s. Our ARFIMA estimations strongly point out towards d ≈ 05, which is exactly the value predicted by canonical SOC models. Therefore, when explaining Brazilian growth cycles, we cannot reject the viability of this theory, which implies, as a corollary, that output fluctuations are in a large extent unavoidable.