The latest issue of the Journal of Economic Perspectives had a very interesting symposium on the costs and benefits of the various bailouts implemented during the Great Recession.
The first article in the issue is by University of Chicago Professor Austan Goolsbee and Princeton Professor Alan Krueger. Goolsbee served on the President’s Council of Economic Advisers from March 2009 to August 2011, and Krueger spent much of 2009 to 2013 in the Treasury Department and CEA, so one might not expect them to be big critics of the policies. Their review is quite candid in communicating the misgivings that many of those in government had about the measures. Here’s Goolsbee and Krueger’s summary of the bottom line:
The US Treasury recovered a total of $39.7 billion from its investment of $51.0 billion in GM. By the end of 2014, Treasury sold its remaining stake in Ally Financial, recovering $19.6 billion from the original $17.2 billion investment in Ally, for a $2.4 billion gain for taxpayers. In May 2011, Chrysler repaid its outstanding loans from the Troubled Asset Relief Program (TARP) six years ahead of schedule. Chrysler returned $11.2 billion of the $12.5 billion it received through principal repayments, interest, and cancelled commitments, and the Treasury fully exited its connection with Chrysler.
And what did any of the rest of us gain from this?
Since bottoming at 623,300 jobs at the trough of the recession in June 2009, employment in the motor vehicles and parts manufacturing industry has increased by 256,000 jobs (as of July 2014). This is a stark contrast from the previous recovery, when jobs in the industry steadily declined. The increase in the number of jobs in motor vehicles and parts manufacturing accounted for nearly 60 percent of the total rise in manufacturing jobs in the recovery’s first five years. In addition, some 225,000 jobs have been added at motor vehicle and parts dealers. Counting both manufacturers and dealers, auto-related jobs accounted for 6 percent of the total 8.1 million jobs that were added, on net, in the first five years of the recovery—triple the sector’s 2 percent share of total employment.
The authors conclude:
It is fair to say that no one involved in the decision to rescue and restructure General Motors and Chrysler ever wanted to be in the position of bailing out failed companies or having the government own a majority stake in a major private company. We are both thrilled and relieved with the result: the automakers got back on their feet, which helped the recovery of the US economy. Indeed, the auto industry’s outsized contribution to the economic recovery has been one of the unexpected consequences of the government intervention.
The symposium also features an analysis of assistance to financial institutions under the Troubled Asset Relief Program (TARP) by Professord Charles Calomiris and Urooj Khan, both at Columbia. Calomiris is a prominent conservative economist, so one might have expected this review to be somewhat critical of the intervention. But the authors acknowledge the favorable bottom line:
For the most part, the [TARP] transactions with the banks, the focus of this paper, yielded a net cash flow gain. The net cash flow costs were largely from the assistance provided to AIG, the automotive industry, and the programs aimed at avoiding home mortgage foreclosures. The net cash flow gain estimated for the Cash Purchase Program was $16 billion with only $2 billion of preferred stock remaining outstanding. The CBO estimated a net cost of $15 billion to the Treasury for the assistance provided to AIG under the Systemically Significant Failing Institutions program. All of the supplementary support provided to Citigroup and Bank of America through the Targeted Investment Program had been paid back and resulted in a net gain of roughly $4 billion dollars to the federal government. Finally, the loss-sharing agreement with Citigroup through the Asset Guarantee Program yielded a net gain of $3.9 billion.
The authors review evidence for the benefits of these programs and find it difficult to make a definitive statement:
Following the announcement of the Capital Purchase Program on October 14, 2008, the first program of TARP announced in the pre-election phase, there were broad improvements in the credit markets. Between Friday, October 10 and Tuesday, October 14, the Standard and Poor’s 500 rose by 11 percent and the common stock prices of the nine large financial institutions that were the very first participants of TARP increased by 34 percent (Veronesi and Zingales 2010). From October 13, 2008 (before the announcement of the CPP) to September 30, 2009, the LIBOR rate fell by 446 basis points and TED spread fell by 434 basis points. Costs of credit and perceptions of risk declined significantly in corporate debt markets as well. By the end of September 2009, the Baa bond rate and spread had fallen by 263 and 205 basis points, respectively (US GAO 2009, p. 37)….
Using an event study analysis of bank enterprise values, Veronesi and Zingales (2010) analyze the effect of the initial announcement of TARP assistance to the financial sector. They estimate that the October 13, 2008, announcement resulted in a net social benefit to financial intermediaries, after subtracting the cost to taxpayers, of between $86 billion to $109 billion….
[But] Veronesi and Zingales (2010) underestimate the expected costs of TARP as of October 13, 2008. The first round of assistance provided to the big banks effectively committed the government to a “whatever it takes” approach to keep AIG, Citigroup, and Bank of America alive, and therefore, the continuing cost to taxpayers actually experienced in 2008–2012 was predictable, at least to some degree. In other words, if TARP assistance would be forthcoming (and more junior in form over time) in response to worsening bank condition, the recipients effectively possessed a put option from the government to issue equity in addition to the explicitly recognized preferred stock investments made by the government….
With regard to TARP’s gross benefits, a credible evaluation of the impact of TARP assistance to financial institutions remains elusive. First, it is difficult, if not impossible, to isolate the effects of TARP from other initiatives of the Federal Reserve, Federal Deposit Insurance Corporation, and other financial regulators, or from other influences on the economy unrelated to government programs. For example, on October 14, 2008, the Capital Purchase Program was announced jointly with the Fed’s Commercial Paper Funding Facility Program and FDIC’s Temporary Liquidity Guarantee Program. Furthermore, it is hard to know to what extent the financial markets would have stabilized and the economy would have recovered in the absence of an activist government response. Some have argued that government support for financial institutions during the crisis confused and frightened market participants and was itself possibly a net negative for the economy.
The Journal of Economic Perspectives symposium on the bailouts also included an analysis of the federal government assumption of conservatorship of mortgage giants Fannie Mae and Freddie Mac by Scott Frame of the Federal Reserve Bank of Atlanta and Andreas Fuster, Joseph Tracy, and James Vickery of the Federal Reserve Bank of New York. Here again the bottom line so far has been a net financial gain for the government:
As of end-2014, the cumulative Treasury dividend payments by Fannie Mae and Freddie Mac have now exceeded their draws: specifically, Fannie Mae has paid $134.5 billion in dividends in comparison to $116.1 billion in draws, while Freddie Mac has paid $91.0 billion in dividends in comparison to $71.3 billion in draws.
And what did ordinary citizens gain from this?
Was it important to promote mortgage supply during this period given the already high levels of outstanding US mortgage debt? We would argue “yes,” for two reasons. First, mortgage origination was necessary to enable refinancing of existing mortgages…. Second, continued mortgage supply enabled at least some households to make home purchases during a period of extreme weakness in the housing market.
And I have separately examined the Federal Reserve’s participation in emergency lending and concluded that the Fed came out making a profit from its intervention as well.
Like Calomiris and Khan, Frame and coauthors suggest that a positive net cash flow is not the right metric:
As an economic matter, one cannot simply compare nominal cash flows but must also take into account that the Treasury took on enormous risk when rescuing [Fannie and Freddie] in 2008 and should therefore earn a substantial risk premium, similar to what private investors would have required at the time, in addition to the regular required return.
Here I disagree. A key argument for intervention was that private risk premia at the time were too high. Prices of risky assets and the volume of risky lending were depressed by fire sales and a scramble for liquidity and safety, which posed a danger of pushing the economy deeper into crisis. Yes, the government was assuming risk in all these actions, risks that ultimately would have been borne by taxpayers. But taxpayers also faced a very real risk of falling revenues associated with a worsening economic downturn. The contributors of the symposium were all correct in emphasizing the risks the government took in making the bailouts, and I agree with their warnings that we should only enter into such ventures with great reluctance and careful evaluation of the costs and alternatives. But I think an objective observer would conclude that this time, at least, it all turned out well.
It seems, consumers paid back taxpayers, e.g. through more and higher fees, which contributed to the slow recovery.
The Bailout Scorecard – Last update: Jun. 2, 2015
There is a big contrast between the automaker bailouts and the bank bailouts. The first put hundreds of thousands of people back to work. The second put a handful of millionaire bankers back to work.
The difference is in the approaches Bush took for the bank bailouts compared to Obama for the auto industry. In Obama’s case he took over the failed auto companies in bankruptcy, wiped out the stockholders, gave the bondholders a haircut, fired and replaced all the failed top management and then sold the company back to the public. For the banks, we have the same incompetents and criminals running the show as before and the country is worse off as a result.
Well, the economy didn’t absolutely need the auto bailouts. For Obama, it was more making sure the recession didn’t linger into his 2012 campaign after the banking bailouts. For Bush, it was the total collapse of the economic system and middle/upper class white people being shutout of their banks, all liquidity gone, would have caused a firestorm. The run on the dollar would have been the final stage as foreigners would bail on the once proud economy. All that was left was citizens demanding federalization of banking and investment leading toward a reborn statist nationalism.
I think a total collapse would have been funny theatre as the white/black divide with unemployment would have closed rapidly by the early 2010’s and tribal wars would have began for the scarcity of resources.
Joseph, you may be unaware of a few facts.
Citigroup alone has 239,000 full time employees.
And, Citigroup stock, in the crisis, had a crew cut, falling to 97 cents a share.
The U.S. government lost billions of dollars in the auto industry rescue (people pay one way or another).
GM, for example, was in decline for decades before it went bankrupt.
And, bankers/top management are no more incompetent or criminal than anyone else (although they tend to work harder) 🙂
I’m quite aware of the facts.
None of Citigroup’s tellers needed to lose their jobs. It is their incompetent managers who should have lost their jobs. The government should have taken over the banks just like they took over GM — wiped out the shareholders, cut the bondholders, fired and replaced the management, and sold the banks back to the public. The hundreds of thousands of bank employees would have continued as before.
“GM was in decline for decades before it went bankrupt.”
Well you demanded facts and the fact is that GM had the largest worldwide sales of any automaker for a string of 77 years from 1931 to 2007. In 2007 they lost the lead to Toyota, but regained the lead in 2012. Since then GM and Toyota have been neck and neck. Hardly a failed company. The incompetent management got into financial trouble because they failed to properly fund their pension and health funds, instead looting the company for excess dividends and management compensation.
“And, bankers/top management are no more incompetent or criminal than anyone else.”
Well, that is certainly debatable. The banks have paid over $200 Billion dollars in fines and penalties for, among others:
Price fixing electricity markets.
Price fixing municipal bond debt.
Price fixing LIBOR.
Price fixing exchange rates.
Price fixing commodities futures.
Price fixing credit card fees.
Falsifying foreclosure documents.
Falsifying mortgage backed securities.
Falsifying debt collections.
Credit card fraud.
Bribing foreign government officials.
Participation in ponzi schemes.
Illegally foreclosing on in-service military troops.
Auto finance fraud.
And that’s just one bank, JP Morgan. To that you could add HSBC — drug trafficking and terrorist financing. Is there any other industry that has come close to that level of criminality?
Joseph, what about GM paying each union worker much more in compensation than other automakers. Was that management’s fault too?
GM’s market was in decline, since the 1960s. Over the “long boom,” its market share fell from roughly 43% in 1982 to 22% in 2007.
Government can easily squeeze money out of banks, even while heavily regulated. After-all, they have deep pockets, which lawyers love.
I agree, it’s difficult for a CEO to micromanage and know everything going on in a huge global bank.
Maybe, we should replace people with robots to eliminate crimes.
No, there is not. Except the dedicated one’s, like the mafia or cosa nostra.
Have worked long enough for the banking industry.
Hamilton is absolutely correct to criticize the last argument about government rescues and risk premia. There is a second problem with their argument that he hasn’t mentioned though. A large part of why the premiums on Fannie and Freddie were high was because of their vulnerability to a self-fulfilling crisis. The economic size and strength of the government means it doesn’t have to worry about that.
It’s hard to argue with JDH’s broader point, but I think it is possible to be a little more specific. As Kenneth Arrow argued, private agents should be risk averse, but government should be risk neutral. In other words, government should assume risks that the private sector wouldn’t touch. The main reason for the auto bailout was a concern that it would spread across the economy and bring down healthy companies. So in evaluating the costs of the bailout you have to consider the costs of not bailing out the auto industry. Same holds true with the bank bailouts. I’m with Joseph on this one. Bailing out the banks was necessary at the time, but there was nothing necessary about the banker friendly way in which the Bush Administration went about doing it. Why are the same crooks in charge today who were in charge of the banks 7 years ago still in charge today? They belong in prison. What other industry pays hundreds of billions in fines and still keeps the same management team at the top?
I stated in Feb 2009:
“Instead of loans for the auto industry, the government should buy autos and give them away to government employees (e.g. a fringe benefit). So, automakers can continue to produce, instead of shutting down their plants for a month. Auto producers should take advantage of lower costs for raw materials and energy, and generate a multiplier effect in related industries.”
In Feb 2009, GM laid-off about 20% of its domestic workforce – 50,000 people – and a total of 107,000 layoffs in the recession. GM domestic auto sales fell from about 4 million in 2007 to a little more than 2 million in 2009. GM’s cost-cutting still couldn’t make it solvent.
Taxpayers lost roughly $10 billion to $15 billion (depending on the accounting) bailing-out U.S. automakers. If the government bought $20 billion of cars from GM instead (e.g. 1 million $20,000 cars, although GM was best producing SUVs, pick-ups, and luxury cars), there would’ve been much fewer layoffs and taxpayers would have something to show for the bailout.
In WWII, government bought jeeps, tanks, ships, etc., which put Americans to work, although production went to war rather than to consumers.
Or, instead of giving away autos, the government could buy autos and give vouchers to dealerships. So, consumers can buy autos made in the USA at a discount for a limited time.
It could be a one-time event, like WWII (although, there was also WWI).
James, nice post.
Next GDP quarters will be shiny, and Hillary will smile even more. Although she is a whore to corporate America and mostly interested in looting the system, esp. given who her financial backers are, she should be in the same clown car as the useless Republican candidates.
I just love it when experts proceed to assert benefits based on the analysis of small balance sheets not encompassing the range of the costs and benefits. Cash flows to the US Govt indeed.
Add in all the other balance sheets, not just the automakers, but all the families and other entities.
The bailouts may have been NPV positive in aggregate (likely), but NPV positive to only some balance sheets, mostly the largest.
Meanwhile it was possible to design fiscal measures for primary benefit of the small and secondary benefit to the large to much greater NPV.
“The US Treasury recovered a total of $39.7 billion from its investment of $51.0 billion in GM. By the end of 2014, Treasury sold its remaining stake in Ally Financial, recovering $19.6 billion from the original $17.2 billion investment in Ally, for a $2.4 billion gain for taxpayers. In May 2011, Chrysler repaid its outstanding loans from the Troubled Asset Relief Program (TARP) six years ahead of schedule. Chrysler returned $11.2 billion of the $12.5 billion it received through principal repayments, interest, and cancelled commitments, and the Treasury fully exited its connection with Chrysler.”
Do the above summaries of “net cash flow” include the financing costs to Treasury? Very few of the reports I’ve seen on this subject do. Never mind risk premia, these studies seem to assume that the financing costs are zero. Sure, borrowing rates for the Treasury are very low, but they are not zero.
Federal government lost $9.26B on auto rescue
December 30, 2014
“Under government accounting rules, the U.S. Treasury actually lost $16.56 billion on paper on the auto bailout…taxpayers lost more because interest and dividends paid by borrowers — in this case, the automakers and finance companies — aren’t applied toward the principal owed.
A homeowner, for example, who borrows $100,000 doesn’t get credited with interest payments in paying off the mortgage. That largely explains the difference between the government’s larger accounting loss and the $9.26 billion net loss.”
That’s a ridiculous analysis from the Detroit paper. So, interest and dividends are not counted against principal? How shocking!
And, it doesn’t address my point. Interest (and dividends) *should* be counted to determine if the US has experienced a net gain or loss on these transactions. However, so should its borrowing costs. Imagine, to take the mortgage example, if a bank were to determine its profit and losses on lending activities by considering only the principal and interest payments it receives from its mortgage borrowers, but ignores the interest expense it pays to the sources of the banks financing, for example, its depositors and bond holders. These costs are likely ignored because the Treasury’s borrowing is fungible. There is no specific Treasury bill or bond issuance one can point to in order to calculate the borrowing costs. However, one should take into account the average cost of borrowing on Treasury debt. Again, this is more fundamental than risk premia.
Under government accounting, interest and dividends aren’t counted. So, the government had a bigger loss.
It should be noted, interest and dividends are taxable.
Anyway, interest rates on bailout loans aren’t important. What’s important is stopping a severe downturn, prevent massive bankruptcies, raise stock prices, and start getting people back to work (to create taxpayers instead of unemployment spending).
“…Interest and dividend are taxable”—–to the government?
Come on, this isn’t so hard. The Hamilton post and the studies he cites refer generally to “net cash flow”. What we are trying to get out of this is whether the government incurred a net cash flow gain or loss on these deals. My point, which is very basic, is that these studies generally don’t count the government’s borrowing costs in calculating the “net cash flow”. But, they should. The cost of providing those funds was not zero because the Treasury obtains the funds by borrowing.
prof hamilton, very nice post. it is amazing the number of people who opposed even the idea of the government bailouts. you could tell a persons perspective simply from the tribe they associated with.
it really is a bit scary to think somebody like mitt romney could have been making the decisions in this time of trials and tribulations. let gm fail was his plan of action-he said it, not me. he failed to understand how large and interconnected the financial crisis was at the time, because he was obsessed with his “free market” interpretation of the economy. and he was right, of course, that gm needed to fail and be reorganized. but what he failed to understand, at that time, there was no private enterprise in existence with the deep pockets of capital needed to move gm into and out of bankruptcy, especially since all of the supporting industries of the automaker would have failed as well. sometimes the plans of a private equity investor cannot be scaled to the economy as a whole.
Prof Hamilton: When a commenter calls a woman a whore I think he/she should be banned from comments.
One could argue just as well that the rebound in American purchases of cars proved that the bailouts weren’t needed. The unstated assumption that seems to run through all these assessments is that the US federal government was the only single entity that could keep GM and Chrysler operating. So therefore they assume that without the bailouts the two remaining domestically owned US carmakers would have shut down, all their productive assets would have been permanently idled, and the recovery of US demand for autos would have been met by foreign car and truck production and thus also foreign parts. The mention of the rebound of dealership jobs as a consequence of the bailout seems to reflect a simple logic error.
If these economists want to enter an informed discussion about the consequences of the auto bailouts, they need to learn much more about bankruptcy and the auto industry. They need to think about (at leat) a) if anyone else might have been willing to inject money and push an accelerated bankruptcy if the US hadn’t signaled willingness, and if so who and on what terms b) how long could a plain vanilla bankruptcy have dragged on, and what would have been the impact on employees and the economy, c) what would have happened to auto worker pensions in a regular bankruptcy, d) who would have acquired GM and Chrysler productive assets in a plain vanilla bankruptcy. Comparisons of reality to a fantasy scenario in which all of GM and Chrysler’s productive assets vanish into thin air (and/or demand for automobiles doesn’t recover) don’t have value.
The unstated assumption that seems to run through all these assessments is that the US federal government was the only single entity that could keep GM and Chrysler operating.
The govt tried to find private investment firms that would take on the risks of GM and Chyrsler. No takers. It’s not an “unstated assumption”; it was a well known fact. The govt was the only entity capable and willing to take on the risks. Remember, a govt’s aversion to risk is and should be very different from a private sector agent’s aversion to risk.
Essentially, Tom is correct. If there are several million consumers who want to buy Chevrolets and Cadillacs, the finance needed will arise to see that they are built. Option 1: GM re-organizes in bankruptcy; shareholders, bondholders and labor all take haircuts and GM continues to produce vehicles. 2: Toyota, Nissan, Honda etc…buy up the assets of GM and produce those cars (under new ownership). 3. Toyota et al. expand their own production lines and supplant GM’s products–i.e. fewer, or none, Chevies get built, but more Corollas, Maximas and Accords do. Not rocket science, as one used to say.
Which is a completely DIFFERENT situation for the fractional reserve banking system which was on the verge of what was allowed to happen in 1930. Back then the banks were allowed to fail, taking something like 1/3 of the money supply out of circulation and producing a huge deflationary spiral. As Calomiris might have put it (actually, he and Stephen Haber did), the latter situation is fragile by design. And it is unlikely that ‘the market’ will resolve the problem quickly and painlessly–well, will the pain kept to a minimum.
Unlike the auto manufacturing issue, which is a conventional microeconomic one.
tom, you are making the same mistake of mitt romney. this was a classic situation of a free market failure. there were no financiers to move gm into and out of bankruptcy. they did not exist. the market failed. why? because everybody was in financial turmoil. it was not just an isolated company struggling-the world economy was struggling. free markets do not work very efficiently in such situations. unless you are accepting of an outcome of chaos and anarchy?
Also, it’s a pretty funny idea that the wisdom of public bailouts should be judged on their capital gain or loss. If that were true then we should be criticizing the government for not buying shares in other sectors. Virtually any random portfolio of equity investment made at that time would have performed far better.
tom, people only use the capital gain loss as an example of why the bailout was not a waste of money. it is in response to people who arbitrarily rate any public bailout as a failure on principal.
“Virtually any random portfolio of equity investment made at that time would have performed far better.”
true. but then again, that is only because nobody was investing in equity at the time due to the severe financial conditions and trepidation of the future. or stated differently, gm should have been a slam dunk investment with money tearing down the walls to own its equity. but apparently i missed that episode in the crisis. uncertainty at the time placed nearly all market participants on the sidelines.
Is this a case of applying accounting and economic analysis to a political calculation? The top 3 employers in Michigan before the recession were GM, Ford, Chrysler (maybe they still are). It was stated at the time that if GM and Chrysler were not bailed out that enough suppliers would be forced into bankruptcy to potentially put Ford into bankruptcy. Reality or just worst case worrying? Who can say. But if you were a new Democratic President and you had the choice of letting the top 3 employers in a blue state do down…
Let’s see here. They spend $12B and saved 250k jobs, which works out as $48,000 per job.
Given that these people probably paid $12k per year in taxes and did not collect
$18k per year in unemployment insurance, that means that the whole thing
paid for itself in a little over two years.
Banks and other financial entities who did not want TARP funding were forced to take it. Many paid the money back as soon as it was legally possible. That distorts any objective analysis of the program because the implication is that repayment and return were positive when those who were forced to take the money were forced to take a lesser alternative in their view. Those who did not want TARP funding believed that they would have gotten a better return but were not allowed.
Also the damage that TARP did to the economy buy being the catalyst that precipitated the market decline that began immediately on the passage of TARP and continued for over 6 months is never considered as part of the cost.
Finally, history has shown that when bloated companies go bankrupt and then are acquired by stronger firms the bloated companies turn around and become extremely profitable. In the case of the bailouts of sick companies, the opportunity costs of companies being run as lean successful businesses is never considered. The only analysis is the seen and the unseen is not considered (Bastiat).
“Banks and other financial entities who did not want TARP funding were forced to take it.”
that was the price to pay for having the federal government act as a backstop against any bank runs. there was a cost for that protection. and ALL of the banks needed it at the time. any bank which existed without such a backstop would have failed. nobody would have kept assets in a bank without such protections.
“Also the damage that TARP did to the economy buy being the catalyst that precipitated the market decline that began immediately on the passage of TARP and continued for over 6 months is never considered as part of the cost.”
you lose all credibility when you argue that TARP was the cause of the market decline. foolish statements. the market had already declined over 20% from its peak when the first tarp bill was rejected by congress. the the market collapse was already in progress when tarp was passed. let me reiterate-the market collapse was already underway. please don’t make such foolish statements.