This paper surveys hundreds of published social-science studies of private, hard-core cartels that contain 699 observations of long-run overcharges. The primary finding is that the median cartel overcharge for all types of cartels over all time periods is 25%: 19% for domestic cartels, 32% for international cartels, and 31% for all successful cartels. Thus, international cartels have historically been about 68% more effective in raising prices than domestic cartels. Cartel overcharges are skewed to the high side, pushing the mean overcharge for all types of cartels over all time periods to 42%. "Peak" cartel overcharges are typically double those of the long-run averages. These results are generally consistent with the few, more limited, previously published works that survey cartel overcharges. There is no evidence that convicted cartels are markedly less effective than unpunished ones. The results of a second survey of final verdicts in decided U.S. horizontal collusion cases, only three of which were international cartels, show an average median overcharge of 21% and an average mean overcharge of 30%. Outside the United States, 62 decisions of competition commissions cited median average overcharges of 25% and a mean of 47%. There are three significant policy implications. First, there is a view among some antitrust writers that there is little evidence that cartels raise prices significantly for a period long enough to justify the height of current U.S. cartel penalties. This survey's results, which are based upon an extraordinarily large amount of data spanning a broad swath of history of all types of private cartels, sharply contradict these views. In fact, the data suggest that U.S. penalties ought to be increased. Mean overcharges are three times as high as the level presumed by the U.S. Sentencing Commission. Surprisingly, bid rigging was no more injurious than other forms of collusion, which suggests that the USSC should amend its Guidelines that currently treat bid rigging more harshly than other forms of collusion. Second, the principal antitrust authorities abroad often base their typical or maximum fines on a 10% harm presumption. Average fines imposed since 1995 by Canada and the EU on identical cartels have been lower than U.S. government fines, yet overcharges generated by cartels discovered outside the United States are higher than North America-centered cartels. Consequently, anticartel laws and fine-setting practices abroad are in even greater need of strengthening. Third, cartels with multi-continental price effects are the most harmful type. Despite the evident increases in cartel detection rates and the size of monetary fines and penalties in the past decade, a good case can be made that current global anticartel regimes are under-deterring. While the recent worldwide trend towards the intensification of cartel penalties has been desirable, global cartels are more difficult to detect, have less fear from entry of rivals, achieve higher levels of sales and profitability, and systematically receive weaker corporate sanctions than comparable domestic cartels. Antitrust sanctions worldwide should be higher for global cartels than for other types.