Liberty Street Economics at the Fed NY has an interesting article on discount window (DW) stigma:
Although it discouraged DW borrowing, the Fed generally kept the DW rate below the market rate, in part because the Fed lacked independence from the Treasury and was obliged to keep the DW rate below the market rate to help the federal government finance its deficits at low rates. The Treasury–Federal Reserve Accord of 1951 freed the Fed from pressure from the Treasury, but the Fed continued to maintain the DW rate below the market rate despite recommendations to the contrary. It did so because it believed that banks that legitimately needed DW funds should not face a punitive rate. Thus, between 1914 and 2003, the DW rate was generally below the market rate on banks’ primary sources for short-term funding (in other words, the commercial paper rate before 1954 and the federal funds rate since 1954; see chart below).
I have come across many economists who believe that the interest rate on treasury bonds is market-determined. In my forthcoming book I explain why the short-term interest rate set by the central bank has a very strong influence on the interest rates and yields of government bonds (and notes). Here you hear it from the Fed NY. Supply and demand for government bonds do not determine the interest rate the government pays. Therefore, loanable funds theory is plain wrong and so is the IS/LM model with crowding out and liquidity trap.
The article contains another clarification which I also describe in my forthcoming book – banks can borrow required reserves if they don’t have sufficient reserves:
Indeed, these requirements may have led market participants to presume that if a bank was borrowing from the DW, it must be in trouble, even if, in fact, the bank was borrowing to address a temporary funding shortfall or to meet reserve requirements.
This means that reserves requirements have to be fulfilled ex-post, being calculated after lending has occurred. Hence a lack of reserves cannot stop banks from lending.