A number of comments to Joe Nocera's September 20 NYT op-ed commentary "No Extra Credit" regarding my hypothesis that a lack of bank credit creation is the principal factor holding back the current economic recovery was that I had missed the point -- it's demand, stupid, or more precisely, lack of demand. If there were more demand for goods and services in the economy, then corporations allegedly sitting on all that cash would start to use it to hire more people. I agree with these comments. Our current weak economic growth is largely the result of inadequate aggregate demand for goods and services, not inadequate aggregate supply. And that is why I suggested a properly designed Federal Reserve quantitative easing could "chum up" aggregate demand until banks are able to create adequate amounts of credit on their own to get the job done. Monetary policy is all about affecting aggregate demand; fiscal policy is all about affecting aggregate supply.

As former Governor Sarah Palin noted this past November and current Governor Rick Perry alluded to this past August, the Fed can, figuratively, print money. And, when this "freshly-printed" money finds its way into the banking system, an adequately capitalized banking system, in turn, can create some credit for the economy figuratively "out of thin air." Credit created out of thin air has the unique characteristic of allowing the recipients of this credit to spend while not requiring the grantors of this credit to cut back on their current spending. This implies that an increase in credit created out of thin air will lead to a net increase in spending in the economy. This is why there is a high positive correlation between percentage changes in bank credit and percentage changes in the domestic demand for goods and services. For a moment, imagine an America without a banking system, but with a Federal Reserve. Although the Federal Reserve Act might not permit it, in theory, the Fed could grant home mortgages to you and me just as banks would, if they existed. In this case, the Fed would be creating credit out of thin air just as the banking system normally does. Okay, snap out of it. We do have banks. But it appears as though a lot of our banks are unable to grant you and me mortgages today. In theory, then, the Fed could temporarily take over this function until banks were again in a position to resume their normal lending role. Because this would make Governors Palin and Perry apoplectic, it, of course, will not be done. But until more banks are in a position to start lending to us and, for that matter, lending more to the government, the Fed could indirectly create more credit for us so that are demand for goods and services would increase.

To that end, I suggested that the Fed target some specified rate of growth in combined Federal Reserve and bank credit. Just for the sake of argument, assume that 5% were the "correct" rate of growth in this combined credit aggregate. If banks on their own were not creating credit growth of 5%, the Fed could engage in some quantitative easing by purchasing securities from the public such that combined Fed and bank credit grew at 5%. How could this increase the aggregate demand for goods and services? Let's go through several scenarios.

In the first scenario, let's assume that the ultimate seller of the security to the Fed is a pension fund. The pension fund now has more cash and fewer interest-bearing securities. From where did the pension fund's extra cash come? Was it from my account? No. Was it from your account? No. The new cash appeared like manna from Heaven. That is, the Fed created it "out of thin air." So, the Fed has created net new cash in the economy. Unless the pension fund now desires to hold more cash and fewer securities with an explicit rate of return, it will look around for other securities to purchase to replace the ones it just sold to the Fed. So, the pension fund will start bidding up the price (and consequently, down the yield), at the margin, of other securities. At the lower yield, the quantity of credit demanded by some entity, a household and/or a business, will increase. That is, some entity will issue a new security in order to obtain funds to purchase something. Let's assume that at the lower interest rate, you and I decide to take out mortgages in order to snap up some these current bargains in residential real estate. The mortgage banker sells our mortgage to Wall Street for securitization. Our mortgages get securitized and the pension fund buys them. This will get the aggregate demand ball rolling. Net new credit will be created in the economy, enabling the recipient of this credit to increase its current spending without requiring any other entity to cut its current spending.

In another scenario, suppose the pension fund, rather than replacing the securities it sold to the Fed, purchases some already existing securities. That is, suppose the pension fund purchases some securities issued in the past that funded some purchases that occurred in the past. How will this get the aggregate demand ball rolling? It depends on what the seller of securities to the pension funds does with its new cash. It might spend it. It might lend it to an entity that will spend it. In either of these cases, the aggregate demand ball will get rolling. Only if the seller of "old" securities to the pension fund desires to just hold more cash will the aggregate demand ball not get rolling.

In yet another scenario, assume that the Fed purchases securities not from a pension fund, but from a bank. If the bank replaces the securities it sold to the Fed with some other securities or loans, then we have the equivalent of the Fed securities purchase from the pension fund. Net new credit will be created in the economy enabling the recipient of this credit to increase its current spending without requiring any other entity to cut its current spending.

But what if the bank chooses not to replace the security it sold to the Fed, but to just lend the new funds back to the Fed now that the Federal Reserve pays interest on banks' cash? This would mean that bank credit would fall. Remember, the quantitative easing approach I am recommending is that the Fed target the growth in combined Federal Reserve and commercial bank credit. If bank credit goes down by the amount that Fed credit goes up, then the Fed would have to engage in even more securities purchases until this combined credit grows at the specified rate. Given that banks already hold over $1.5 trillion of excess cash (excess reserves), it is unlikely that banks would be interested in selling securities to the Fed just to top off their excess cash.

What if Paul Krugman is correct that the demand for cash is now infinitely elastic with respect to interest rates? That is, no matter how much new cash the Fed creates out of thin air, the public, including banks, will be content to passively hold it without using any of it to increase spending. If this is the case, then the quantitative monetary policy I have recommended would not lead to an increase in aggregate demand. But because it couldn't hurt aggregate demand, why not give it a proper shot?

This notion of an infinite interest elasticity of demand for cash, the Hicksian liquidity trap, came into vogue in the 1930s. But as I showed in "If Some Dare Call It Treason, Was Milton Friedman a Traitor?," when bank credit revived in 1934, so, too, did nominal GDP growth (see chart below). Evidently, liquidity was not such a trap in the 1930s to prevent growth in bank credit from reviving aggregate demand. If liquidity did not prove to be so potent a trap in the early 1930s to prevent increases in bank credit from stimulating aggregate demand, liquidity demand would not now seem to be inescapable from with increases in Fed and/or bank credit.

By the way, what caused bank credit to resume growth in 1934? Two things. Firstly, after FDR declared the nationwide banking holiday on March 6, 1933, two days after his first presidential inauguration, the Reconstruction Finance Corporation engaged in a TARP-like operation, injecting new capital into many banks. Secondly, the establishment of FDIC in June 1933 alleviated runs on banks by depositors. With additional capital and less need for cash reserves to meet deposit withdrawals, banks began adding to their earning assets in 1934, primarily their government securities holdings. Loans on the books of banks did not start to increase until the second half of 1935, and then only marginally. The composition of bank credit is less important than the quantity of total bank credit with regard to aggregate demand.

Paul Kasriel is the recipient of the 2006 Lawrence R. Klein Award for Blue Chip Forecasting Accuracy