We measure the size of the fiscal multiplier using a heterogeneous agents model with incomplete markets, capital and rigid prices and wages. This environment captures all elements that are considered essential for a quantitative analysis. First, output is (partially) demand determined due to pricing frictions in product and labor markets, so that a fiscal stimulus increases aggregate demand. Second, incomplete markets deliver a realistic distribution of the marginal propensity to consume across the population, whereas all households counterfactually behave according to the permanent income hypothesis if markets are complete. Here, poor households feature high MPCs and thus tend to spend a large fraction of the additional income that arises as a result of a fiscal stimulus, assigning a quantitatively important role to the standard textbook Keynesian cross logic. Interestingly, and unlike conventional wisdom would suggest, our dynamic forward looking model reinforces this channel significantly. Third, the model features a realistic wealth to income ratio since we allow two assets, government bonds and capital. We find that market incompleteness plays the key role in determining the size of the fiscal multiplier, which is about 1.5 if deficit financed and about 0.6 if tax financed. Surprisingly, the size of fiscal multiplier remains similar in the Great recession where the economy was in a liquidity trap. Finally, we elucidate the differences between our heterogeneous-agent incomplete-markets model to those featuring complete markets or hand-to-mouth consumers. The ADEMU Working Paper Series is being supported by the European Commission Horizon 2020 European Union funding for Research & Innovation, grant agreement No 649396.