This paper exploits a quasi-natural experiment to study the effect of social benefits level on take-up rates. We find that households who are eligible for double benefits (twins) have much higher take-up rate - up to double - as compared to a control group of households. Our estimated effect of benefits level is much higher relative to the standard cross section estimates. This finding is less exposed to a selection bias that might plague much of the previous research on the link between benefits level and take-up. It provides strong empirical support for the level of benefits as a key factor in determining take-up rates.