Friedman claimed that despite their non-money nature (in so far as they cannot be directly used to barter for goods and services), government securities, being highly liquid and having some certainty about their prices, are closer to money than private sector securities. He therefore refers to them as "money-like assets". As Harless writes,
"When the government borrows, it is increasing the fraction of 'money and money-like assets' in people’s portfolios. Instead of buying less of the (non-money-like) private-sector assets (to get the fraction of money their portfolios back up), they will buy more such assets – to get the fraction of 'money and money-like assets' back down. Instead of 'crowding out' private investment, public borrowing will 'crowd in' private investment."
Friedman also pointed out that as maturities shorten, the similarity with money itself increases - T-Bills are closer to money, often close substitutes - because they can be "sold quickly and at a reliable price". Further, T-Bills issuance by the government makes people's portfolio's safer and therefore "increases, rather than decreases, the incentive to purchase private sector assets". The net result is that "the more short-term financing the Treasury does, the larger economic stimulus it provides".
Given the present zero-bound nominal interest rate, which makes T-Bills exactly like money, this arguement should carry even more relevance and make such purchases expansionary. Harless makes the interesting point that given the present inter-changeable flexibility between money and the ("money-like") T-Bills issued by the Treasury to finance government debt, Treasury assumes charge of the monetary policy too (in so far as it can control the effective 'money' supply by issuing T-Bills). In other words, when the Treasury issues T-Bills, it not only finances government debt, but also provides an expansionary economic stimulus by "crowding-in" private investment.
There is one thing though about the "crowding in" arguement. Under normal times, without deflationary pressures (when inflationary expectations are not far away) and when government debt is not monetized (ie. money is not printed, and government prefers to raise debt by issuing securities, short or long term), there is only a fixed quantity of money supply available (and there is less likelihood of its monetization, given the imminent expectations of inflation) to purchase private sector assets. To the extent that this cash goes into financing government spending, there is a proportionate decrease in money supply available for the private sector.
However, when inflationary expectations are firmly anchored and deflation has taken hold, T-Bills become perfect substitutes for cash, and the dangers of monetization are accordingly taken care of, the "crowding-in" arguement appears plausible.