[THREAD] SWP 761 by Zoltan Jakab + Michael Kumhof argues the standard economic model of banks as intermediaries of loanable funds is wrong. It misrepresents what banks do + doesn’t fit stylised facts about balance sheets + propagation of financial shocks.
bankofengland.co.uk/working-paper/…
bankofengland.co.uk/working-paper/…
2/ Most economic models assume banks are “intermediaries of loanable funds”(ILF): i.e. savers turn up at branches with funds to deposit, which a bank then lends on to borrowers. In this view banks don’t create credit, they just channel existing funds from savers to borrowers.
3/ In practice, banks can make loans by financing through money creation (FMC): i.e. they can loan money without having any pre-existing savings. The bank just creates a new liability (the new funds in borrowers account) + corresponding asset (the new loan on its lending book).
4/ Authors argue this distinction is crucial + goes beyond just accurately describing banks’ operations. Under ILF, total lending lending adjusts slowly, as deposits have to precede credit creation. Under FMC, credit can be created ex nihilo so lending can move quickly + sharply.
5/ Using an otherwise identical DSGE model, authors show that FMC model can generate much larger + more volatile changes in lending, consistent with the observed data which ILF model cannot. And FMC model generates much bigger amplification of financial shocks than the ILF model.
6/ To generate large and volatile changes in lending in ILF models, one needs to rely on the alternatives of volatility in physical savings, net securities purchases or asset valuations. These alternatives have virtually no support in the data
7/ Authors see paper as important initial step towards build macro models based on FMC rather than ILF. They believe limited role for credit + financial variables in ILF models can mislead policymakers. FMC models could yield important insights on monetary + macropru policy.
