Allan Sloan’s bad math and bad analysis

By Bob Eisenbeis, Cumberland Advisors

A recent column by Allan Sloan and Doris Burke in the Washington Post claims that the distasteful financial bailout not only worked but also generated a profit for the government of at least $40 billion and perhaps as much as $100 billion. Their conclusion is based on their working of the numbers, and the source of the so-called “profit” is the interest that the Fed has earned on the assets acquired through QE 1 and QE 2 and returned to the Treasury. They estimate that a net $102 billion has been returned to the Treasury by the Fed in 2010 and indicate that as much as $85 billion more may be returned this year.

Unfortunately, the authors have played fast and loose with the numbers. They have ignored important unreported costs of the bailout and, most importantly, they have misrepresented the true nature of Federal Reserve transfers of earnings to the Treasury. Because of these problems, their analysis risks becoming part of revisionist history that obscures the true costs of the bailout to the taxpayer. Their work is already being cited in the hearings on FOMC monetary policy of July 13, 2011. Let us try to put forth a more objective take on the numbers and the kinds of analysis that must be done to get a clearer picture of the true costs of the bailout.

Consider first the authors’ treatment of the costs associated with Freddie and Fannie, which they state is the source of the biggest costs of the bailout. They pull numbers from a CBO June 2, 2011 report stating that the government has injected $130 billion on net into those institutions. But this ignores other important costs also contained in that same CBO report. Specifically, the CBO states that fair-market adjustments to the assets and liabilities of Freddie and Fannie expose another $187 billion in unrecognized losses that must be added to the expected costs of their failure. Additionally, the CBO also estimates that the value of projected government subsidies to Freddie and Fannie between 2012-2021 will add another $42 billion to the estimated costs, bringing the likely total to $359 billion. This is far in excess of the $130 billion put forward by the authors and is sufficient to generate a significant “loss” on the bailout. But there is more.

I don’t quibble with some of the other estimates except to note that the low-interest-rate environment that the Federal Reserve has engineered has been a clear subsidy to financial institutions. The value of that subsidy should be considered a cost, but doesn’t appear in anybody’s calculations. Then there is the value of the subsidized support provided through the discount window and other Fed emergency programs. Nor do people consider the fact that institutions have been able to borrow in the Federal Funds market at a rate substantially below the 25-basis-point risk-free rate that can be earned by depositing those funds at the Federal Reserve. Banks are also receiving 25 basis points, risk free, on excess reserves held in deposit accounts at the Federal Reserve.

The real problem with the Sloan-Burke analysis, however, is in their treatment of the Federal Reserve’s excess revenues over costs that are transferred to the Treasury. They suggest that the assets purchased as part of QE 1 and QE 2 have generated additional “profits” of about $102 billion for 2010 and another estimated $55 billion for 2011.

As we have noted in previous commentaries, the peculiarities of government accounting conventions treat these transfers as revenue to the Treasury. Thus, the funds can also be counted towards reducing the deficit. Treating an interest expense as revenue when returned by the Fed to the Treasury is really an accounting shell game. Consider what the Fed has done. It has printed money by issuing one form of government debt, to purchase to purchase other forms of government debt, Treasuries and MBS from government-backed Freddie and Fannie. The government (Treasury) pays interest to another government entity (Federal Reserve) on those assets, and when the funds are transferred back to the Treasury an expense is magically transformed into revenue.

If the Fed purchased those securities directly from the Treasury, which it is legally prohibited from doing, the Fed would be directly monetizing the debt. It is now indirectly monetizing the debt. The revenues from this activity are not profits, nor should they be considered a return on investment from the bailout. Rather they are simply an intergovernmental transfer that results from printing money.

If it were that easy to reduce the deficit, then the Fed could simply buy up as much debt as possible. The Treasury could then use the return of interest payments it makes to the Fed to reduce the deficit to zero. If a little works, then why not do a lot? It is like writing a check to your wife, and when she gifts it back, you count it as income. The process makes Bernie Madoff’s Ponzi scheme look like chump change.

Bob Eisenbeis is Cumberland’s Chief Monetary Economist. Prior to joining Cumberland Advisors he was the Executive Vice President and Director of Research at the Federal Reserve Bank of Atlanta. Bob is presently a member of the U.S. Shadow Financial Regulatory Committee and the Financial Economist Roundtable. His bio is found at www.cumber.com. He may be reached at Bob.Eisenbeis@cumber.com.

Allan Sloan’s response:

I’ve read Bob Eisenbeis’ work for years and generally respect it even when I disagree with him. But his criticism of my recent article in Fortune and the Washington Post about the costs of the financial system bailout to U.S. taxpayers is wrong on both the facts and the numbers.

My article estimates the cash costs and cash income to the federal government — which I’m treating as a proxy for U.S. taxpayers — of the various parts of the bailout, including the relatively minor Troubled Asset Relief Program (the notorious TARP) and the huge bailout of creditors of Fannie Mae and Freddie Mac, aka government-sponsored enterprises, or GSEs.

My Fortune colleague Doris Burke and I include as bailout revenue the profits that the Fed has made on its bailout-related securities portfolio (primarily from the mortgage-backed securities it owns as part of “Quantitative Easing 1” and the Treasury securities it owns as part of “Quantitative Easing 2”) and remitted to the U.S. Treasury.

The difference between the Congressional Budget Office cost number for the GSEs that Bob cites and the number that I’m using doesn’t involve cash, which he knows or should have known. Rather, it consists of the CBO’s estimate of the subsidy that taxpayers have extended the GSEs’ creditors by guaranteeing the debt without getting appropriate financial compensation for the risk to which taxpayers were exposed. It was made shortly after the government nationalized the GSEs in the fall of 2008, near the market trough.

The subsidy cost is the proper way for the CBO to measure GSE expense for federal budget purposes. But in the summer of 2011, it’s not the right way to measure the GSEs’ cost to taxpayers. Throughout my article, I’m talking about taxpayer cash out, and taxpayer cash in. I’m not talking about budget impacts, or risk-adjusted rates of return. Just dollars out and dollars in.

Bob notes, correctly, that the CBO projected a budget cost for the GSEs’ post-nationalization book of business. But he doesn’t mention that the CBO (along with the GSEs’ trustee and the Treasury) predicts that the government’s net cash outlay to support the GSEs will decline in the future. A major reason is that the GSEs’ post-nationalization business (for which, remember, the CBO assessed a budget expense) is running cash-flow positive. I was being conservative (in the accounting sense) by projecting the current $130 billion cash cost as the final cost.

As for the Fed, I’m calculating how much above-previous-trend the Fed has remitted to the Treasury from what it’s earned on its bailout related portfolio, primarily its QE 1 and QE 2 holdings. I’m also estimating how much the Fed will remit in 2011 and 2012. Of course, everyone who’s economically literate knows there are real problems — dollar devaluation, moral hazard, the screwing of prudent savers to rescue imprudent institutions — attached to these policies.

I’ve written about those aspects of the Fed policies many, many times, using lines like, “Thanks for nothing, Uncle Sam” about the impact of zero interest rates on savers. But in this article, I’m not talking my philosophical and political book, which in many ways probably isn’t much different than Bob’s. I’m just showing the numbers.

The Fed has traditionally remitted its surplus profits — consisting primarily of income on what used to be $700 billion to $800 billion of Treasury paper — to the Treasury. I haven’t researched the literature, but I don’t think there’s been much serious objection among rational analysts about counting those remittances as income for federal budget purposes.

Now, the Fed’s portfolio is much bigger. And so are its profits, and its profits remitted to the Treasury. It’s the same way things have been for decades, just with bigger numbers. I’ve estimated the portion of the Fed’s 2007 (when the bailout of the world financial system started) through 2010 remittances attributable to its bailout activities, and have estimated the 2011 and 2012 bailout-related remittances. These aren’t any different than the 2006-and-earlier remittances, just bigger.

All the securities the Fed owns as part of QE1 and QE2 would exist and be paying cash interest at government expense even if the Fed had not adopted either program. Absent QE1 and QE2, taxpayers would get none of those interest payments. Because of QE1 and QE2, taxpayers get the interest generated by the Fed’s holdings. So I consider it legitimate to count those proceeds as bailout-related profits, and I subtract them from the cost of the bailout.

The article doesn’t praise QE2, of which I’ve been dubious all along, and doesn’t suggest (as Bob implies it does) that the Fed should monetize the deficit by purchasing all the debt the Treasury wants to issue.

My assignment was to write an article for a general audience about the taxpayers’ cost of the bailout. Much to my surprise, I came to realize that taxpayers are coming out ahead. Because I’m a journalist rather than an ideologue, I wrote what I found to be the case, not what I expected or wanted to be the case.

Having U.S. taxpayers coming out somewhat ahead in cash is a lot better than having them come out way behind, which many people believe (wrongly) to have been the case. Many of our numbers are estimates, as Doris and I disclose repeatedly in the article and accompanying financial data. But they’re consistent, and they’re honestly arrived at. And we stand by them.

Back to: Surprise! We still believe the big bad bailout is paying off