Liquidity considerations matter for macroeconomics. They help explain asset pricing anomalies, the code termination of asset prices and macroeconomic conditions, and the transmission mechanism of monetary policy. Since liquidity can have different meanings, I will define an asset as illiquid if it can be sold at short notice only for a discounted price, or not at all. Liquid assets have an essential role in economies where credit arrangements that would allow households and firms to finance spending shocks (e.g., consumption or investment opportunities) are not always feasible.
A critical observation from Kiyotaki and Moore (2005) and Lagos (2006) is that not all assets are equally suitable for helping agents face these shocks: some assets are more liquid than others. In Kiyotaki and Moore, agents who hold land and capital can use only a fraction of their capital stock to finance investment opportunities. In Lagos, agents hold risk free bonds and equity, but equity shares can only be used to finance a fraction of their consumption opportunities. While these liquidity differences among assets help explain several macroeconomic phenomena, the differences in liquidity themselves are left unexplained by the proposed theories. In particular, no link is made between the characteristics of an asset, such as its degree of risk, the supply of the asset, to the ease with which it is traded.
The aim of this paper is to provide a monetary theory of asset liquidity one that emphasizes the role of assets in payment arrangements and to explore the implications of the theory for the relationship between assetslintrinsic characteristics and liquidity, and the effects of policy on asset prices and welfare. Following the literature pioneered by Kiyotaki and Wright (1989), this paper considers economies in which trading opportunities between randomly matched, anonymous agents makes the use of some assets essential as means of payment.
I consider an economy where two assets can serve this role, risk free bonds and risky equity. (I also consider fiat money in a later part of the paper.) Without additional frictions beyond the ones that rule out credit arrangements, all forms of wealth are equally good as means of payment, and agents are indifferent between which asset they spend or which they accept. In order to overcome this indeterminacy, I assume that the liquidity differential between bonds and equity stems from an informational asymmetry regarding the fundamental value of the equity. Specifically, agents paying with an asset are better informed about its future performance than agents who receive it, which makes it costly to trade.
A key insight of the theory is that the risk free asset is a strictly preferred means of payment, and equity is partially illiquid. In order to finance their consumption opportunities, individuals deplete their bond holdings first, and they use equity as a last resort. Moreover, individuals retain a fraction of their equity holdings even when their consumption is ineq ciently low. A major insight of Kiyotaki and Wright (1989) was to show that the acceptability of a good depends on its storage cost as well as other fundamentals (e.g., the pattern of specialization) and beliefs.
In the same vein, I find that the liquidity of the equity, as captured by the quantity of equity agents can sell before experiencing a deterioration in the terms of trade, depends on its dividend process. Equity becomes more illiquid as the dispersion of the dividends across states increases. In the limiting case where the equity has no value in some states, it becomes fully illiquid and, in the absence of risk free assets, trades shut down.
