creation in our set-up, since they prevent banks from issuing fake receipts to expand the
supply of liquidity. In an extension, we contrast the effect of liquidity requirements with
the eff ect of capital (equity) requirements, which can actually enhance banks’ ability
to create liquidity by reducing moral hazard problems between banks and depositors,
making warehousing relatively more efficient. We also extend the model to include a
central bank and argue that a tighter monetary policy does not alway s lead to lower
liquidity creation. We establish conditions under which such a policy can actually
encourage lending by warehouse-banks.
Our model stands in contrast to the contemporary literature on why banks exist.
The only assumption we make on banks is that they have a storage technology. They
have no superior ability to screen or to monitor loans in an environment of asymmetric
information, as in Diamond (1984) and Ramakrishnan and Thakor (1984). Because we
assume that all agents are risk neutral, banks also do not provide better risk sharing for
risk averse depositors as in Bryant (1980) and Diamond and Dybvig (1983). Technolog-
ical and financial developments have diminished informational frictions and provided
alternatives to banks for risk-sharing (see the discussion in Coval and Thakor (2005)).
This s hould have led to a decline in financial intermediaries’ share of output (and cor-
porate profits) in developed economies, but their financial sectors have continued to
grow. This suggests that other forces also determine the demand for banking services;
we suggest that warehousing-type financial services may be one important determinant,
one linked with the very origins of banking. In fact, the largest deposit bank in the
world today, Bank of New York–Mellon, is usually classified as a custodian bank, i.e.
an institution responsible for the safeguarding, or warehousing, of financial assets.
Our paper is also related to the literature on liquidity creation by banks which relies
on the role of intermediaries as providers of liquidity insurance. Important contribu-
tions include Allen and Gale (1998), Allen, Carletti, and Gale (2014), Bryant (1980),
Diamond and Dybvig (1983), and Postlewaite and Vives (1987). In our model, banks
create liquid securities, since they issue receipts that constitute private money, and they
also create aggregate liquidity since more capital is directed into productive investment
with the bank than without. In other words, banks enhance funding liquidity. This
contrasts with the model in Diamond and Dybvig (1983) and other models of that genre
in at least five important respects, discussed below.
First, in Diamond and Dybvig (1983), the bank creates liquid securities, namely
demand-deposit contracts, to raise funds to invest in illiquid projects. However, the
bank does not create aggregate funding liquidity, since no more is invested in the illiq-
uid projects than would be absent the bank. In our model, the bank creates aggregate
funding liquidity precisely because it can create liquid securities—warehouse receipts—
with which to make loans. Since the receipts circulate as a medium of exchange, they
3