$29 Trillion Bailout: Response to Critics

OK, anytime one criticizes the Fed or Wall Street there will be some push-back by the professionals who serve their masters. (By contrast, Barry Ritholtz understood the argument, see here.)  My original piece on Friday got picked up by a number of blogs and generated a lot of hostile responses. I expect that. It is obvious from their comments that many of them did not bother to read the post very carefully, or, if they did, that they stuck to talking points. And it looks like many of them deal in obfuscations that would make Chairman Bernanke proud.

But let us presume they were not hired by the Fed and Goldman and instead assume good intentions. I will have a longish post over at New Economic Perspectives tomorrow that addresses several issues that were not adequately covered in my post on Friday. But here I will deal with the main topic of a number of the comments—which centered around the proper way to measure the Fed’s intervention: stocks or flows.

Several commentators presume I cannot tell the difference between stocks and flows. No long-time reader of this blog or of NEP would be confused about this. I know the difference, and indeed have been using Wynne Godley’s stock-flow consistent approach for a very long time.

As I said on Friday, we can choose any of three different measures to ascertain the size of the Fed’s response. First, we can look at peak lending at an instant; more practically, we could choose end-of-day lending by the Fed. We can measure this by looking at the Fed’s balance sheet, adding across the assets associated with the emergency lending facilities (a Fed loan to a bank shows up on the Fed’s balance sheet as an asset; a Fed purchase of an asset from a bank moves that asset to the Fed’s balance sheet.) This is the measure the Fed chooses, and it comes to a peak of $1.2 trillion on a day in December 2008.

(It will miss any off-balance sheet commitments. For example, if the Fed extends a guarantee that does not get triggered, it will never show up on the Fed’s balance sheet, meaning that a measure that includes only assets on the Fed’s balance sheet will miss some of the ex ante exposure to risk. We’ll ignore that here. Much like Bernanke’s claim that the lending turned out OK—because most loans got paid back—that is 20-20 hindsight and does not count when it comes to ex ante risk.)

OK, the peak instantaneous lending is a useful measure. I have never denied that. The analogy is to the six shot glasses poured by the bartender. If we want to tally up the Fed’s exposure to losses, that is a relevant measure. But note that my piece was not focused on risk to the Fed. To be sure, $1.2 trillion is a big exposure. It merits some concern. To my mind, this is a second order issue, although it is highly probable that Congress will be very concerned with Fed exposure to losses.

The second measure would be more closely related to a flow: peak lending over some period such as a week or month. That gives a good idea of financial system distress in crisis. We can measure the flow over a discrete period of time, and it accumulates to a stock. This is particularly useful when a new facility is created, as a spike of new lending will increase this measure. In the study undertaken by Nicola Matthews and James Felkerson, the period chosen was a week or a month. To be sure, any discrete period chosen is arbitrary. This gives us an idea of how much the bartender has poured for our drunken bankers over a week or a month or a year or a decade or a lifetime. If we are interested in a chronic drinking problem, the instantaneous measure (6 glasses at any point in time) is useless—we need to know how much was drunk over the week or month or other time period of interest. That lets us know how often the Fed was called-upon by our drunks to keep those six glasses full.

And the third measure totals lending over the life-span of each special credit facility. Some of the facilities were short-term, with a quick rise of lending that was temporary; once repaid, the facility disappeared. As such, its cumulative total would add little to the accounting for the size of the bail-out. Other facilities lasted for years; some maintained high levels of lending for week after week and even month after month. That contributes a lot to the bail-out’s size and also indicates particular parts of the financial sector that were troubled for months.

Much like an alcoholic with a drinking problem, some very big Wall Street firms had to drink from the Fed’s trough for extended periods because they were not able to fund on reasonable terms in markets. Surely this information is of interest to Congress and the public. It should not be dismissed with the easy claim that the billions, and even tens of billions of repeated borrowing at the Fed was “temporary” even as it extended over months and even years.

For example, as reported in Felkerson’s paper (now available here) we can look at all three measures of the Primary Dealer Credit Facility (PDCF) that was created on March 16, 2008 in response to the troubles at Bear Stearns.  The PDCF was effectively a “discount window for primary dealers” to ease strains in the repo market by lending reserves on an overnight basis to primary dealers at their initiative.  As Felkerson writes, “PDCF credit was secured by eligible collateral; with haircuts applied to provide the Fed with a degree of protection from risk.  Initial collateral accepted in transactions under the PDCF were investment grade securities” but that was relaxed “to include all forms of securities normally used in private sector repo transactions.”

If we use the cumulative measure, the “PDCF issued 1,376 loans totaling $8,950.99 billion.”  By contrast, the peak weekly amounts outstanding and lent (both occurred on October 1, 2008) were $146.57 billion and $687.74 billion, respectively.

Hence we get three figures: almost $9 trillion as the cumulative number, almost $700 billion as the peak weekly lending, and nearly $150 billion as the peak outstanding number. Take your choice. It is “horses for courses”: the appropriate number chosen depends on the question asked. The smallest number gives an answer to the question “what was the Fed’s peak exposure to losses (assuming the Fed would let the institutions fail without extending even more credit to them)?”; the middle number indicates how much it took to provide meet liquidity demands during the worst week of the crisis, from the point of view of the dealers. And the big number tells us how much the Fed had to intervene over the entire life of the facility to settle markets.

Further, most of that $9 trillion cumulative borrowing can be attributed to just five “drunk” banks, as shown in the following table from Felkerson’s paper:

Five Largest PDCF borrowers, in billions

Source: Federal Reserve

Clearly, these were troubled institutions—two (Merrill and Bear) disappeared as independent banks, Citi came perilously close to the cliff, and Morgan and Bank of America remain in some distress. Doesn’t the cumulative lending by the Fed contribute to our understanding of the depths of their problems?

I emphasize again that we are not trying to mislead anyone, which is why all three estimates are reported in the paper by Felkerson. Unlike the Fed’s memo, we are providing detailed analysis that allows analysts to draw their own conclusions. If we have made mistakes in our calculations, we are open to correction. We will continue to endeavor to expose the half-truths propagated by the Fed and its apologists. Congress and the public have the right to know what the Fed did. We think that it is important to include all three measures of the Fed’s bail-out of Wall Street.

Let me just deal with one other longish and nasty comment, by one who questioned my use of the word “commitments” as the sum of Fed lending, spending (asset purchases) and guarantees. If the commentator had read the Bloomberg response to the Fed’s memo, he would know I am just adopting the terminology used. The Fed had complained about Bloomberg’s supposed claim that it “lent” $7.77 trillion to the big banks; actually, Bloomberg had said “committed”:

“In a March 31, 2009, story, Bloomberg News tallied the potential commitments of the Fed using as sources statements the central bank made and its weekly balance sheet. The amount, $7.77 trillion, was never characterized by Bloomberg as money lent by the Fed, though other commentators have mistakenly used it in that context. Rather, Bloomberg has said that that amount represents what the Fed “lent, spent or committed” or the total of all “guarantees and lending limits.” Bloomberg has been careful to characterize this number as total commitments, not loans that went out the door.”

By contrast, in the work done by Matthews and Felkerson, only facilities actually used are included in the totals—the $29 trillion figure does not include created facilities that were never used. I have used the term “committed” only to include funds lent or used to purchase assets—including toxic waste assets, some of which are still on the Fed’s balance sheet. All of this is very clear in the Felkerson paper.

In coming blogs I will discuss in more detail who got those bailouts. It is not sufficient to claim that hundreds or thousands of borrowers got sweet deals from the Fed—as the memo released by Bernanke does. Where did the majority of the Fed’s lending go? Who benefited most? How much of the Fed’s largesse went to foreign institutions? These are questions that Americans are asking. As always, I appreciate comments, and expect that the “professionals” will be out in full force to continue to sweep under the rug the Fed’s questionable activities.