Wednesday, September 15, 2010

Stimulus more or less? A failure not being acknowledged. PART 3

The economic impact of fiscal stimulus isn’t a partisan issue. Nor is what was done.

As an example of an analysis let’s use “How the Great Recession Was Brought to an End” by Alan S. Blinder and Mark Zandi. It is available free, often cited by media coverage, and, for a quality economic analysis, exceedingly readable. It achieves a number of other criteria that The Hedged Economist would insist upon before recommending any analysis. Most importantly, it is an analysis.

The article can be recommended for its content regardless of its conclusions. As explained below, one doesn’t have to share the methodological biases/assumptions of the authors or their conclusions to benefit from reading the article. The article includes two appendices in addition to the assessment.

A curious aspect of the article is that Appendix A contains some of the best and most important analysis in the article. Much of it focuses on what was actually done and it’s timing relative to what was going on in the economy. It may seem strange that just separating facts from political posturing could be such a valuable service, but it is.

Part of why stating facts is so powerful is that many commentators are either unable or unwilling to separate their biases from reality. They will report things as true because they think they should be true. Sometimes it’s as stupid as calling a loan a bailout or a rip off depending on whether it’s paid back and who gets it. A loan is a loan. Even when a loan is forced on a borrower as happened with some TARP funds, it’s a loan. Even if it is given to a borrower who doesn’t bother to read the terms or can’t read, it’s a loan.

Letting biases sneak in when discussing the future is unavoidable, and one would be naive to think it doesn’t happen to Blinder and Zandi. However, when discussing the past, Blinder and Zandi seem to have successfully presented the facts without much nonsense. The degree it does sneak in is an unavoidable byproduct of their use of a program by program framework for the discussion. But, to any degree it sneaks into their words it is offset by their liberal inclusion of data and the charts, which they use to present the measures taken during the time period.

To illustrate how a biased perception of what happened leads to erroneous conclusions, consider this quote from a front page article from THE WALL STREET JOURNAL no less. It is from a September 11, 2010 article entitled “Tough Bank Rules Coming” about Basel III. The article isn’t available free so the link isn’t attached.

Following a paragraph discussing increased capital requirements, the article states “…potentially forcing banks to take fewer risks…” Now Basel III will force banks to shed some activities that create risk, but, and this is important, increased capital requirements alone have nothing to do with shedding risk.

One doesn't have to be terribly familiar with what is called a "barbell" portfolio strategy to understand that there is a rational way around this potential impact of having more equity on the balance sheet. Forcing banks to hold more of a conservative asset or a conservative liability can be offset by increasing the holding of an offsetting high risk item. The overall risk doesn’t change. In fact, it’s a rational response if one wants to maintain a given overall portfolio return. One just takes on a high risk, high return positions to offset the low risk, low return position; risk and return stay the same. If the reader doesn’t understand why, consult anything on portfolio theory. Or, just consult Angels, entrepreneurs, and diversification: PART 2. The first few paragraphs provide a quick summary of what is known as modern portfolio theory. It’s actually just common sense.

One might argue the Journal quote is either intentional or unintended deception. But, let’s give the authors the benefit of the doubt. Instead of impugning motives of the Journal’s reporters remember that it could easily result from a biased perception of what caused what. Given that assumption, the telling quote is what follows. Again about Basil III, “It comes nearly two years after the chaotic bankruptcy of Lehman Brothers convulsed the global economy and led to taxpayer-funded bailouts…”

Well, for starters, Basil III was planned; maybe the Basil treaty negotiations had begun before Lehman failed. Lehman may or may not have slowed down the negotiations, but they were unrelated events. But that’s a minor and implied failure to identify what caused what.

The big failure to appreciate what caused what is the rest of the quote. The failure of Lehman Brothers didn’t convulse the global economy. Look at the data not the headlines! Even with the limited data in Appendix A, one can see that something convulsed the financial system in August of 2007, WELL BEFORE LEHMAN. That’s one reason why Appendix A is so important.

Just using the data behind Chart 1 of Appendix A illustrates the point. Lehman’s failure preceded a run up in the spread that was only slightly greater than the run up associated with the collapse of the Bear Stern hedge funds, but only slightly (change from about .5 to 2.4 in August verses 1.2 to 3.1 with Lehman). In fact, the run up in this spread after the initial failure of TARP (3.0 to 4.5) wasn’t different in kind. For that matter the late 2007 period of bank turmoil identified by funding problems also wasn’t different in kind.
The important point is that bank funding problems, bank runs, atypical behavior in all sorts of spreads, increased volatility in spreads, the shape of the yield curve, financial flow other than just bank funding problems, and global stock and bond markets all indicated the global economy and financial infrastructure were being convulsed WELL BEFORE LEHMAN.

Now, let’s look at the second half of the quote: “…taxpayer-funded bailouts…” Looking at what actually happened (i.e., the data) reveals that much of the bailouts were actually loans funded by future earnings of the organizations that were lent the money. That’s what every loan is if it is paid back. In addition to loans, and sometimes accompanying the loans, there were guarantees: basically puts. The puts associated with the loans often turned out to be reasonably profitable for the seller of the puts (i.e., the taxpayer); again based on future earnings of the organizations that had to buy the puts. The puts that were given away are a different story.

Not all the puts or loans were profitable nor were they all treated the same. Appendix A of the Blinder and Zandi article sorts some of that out. A September 13, 2010 “Heard of the Street” column entitled “Where Are the AIG Dividends?” by Peter Eavis is also worth reading. It is especially relevant because of how it indirectly highlights the need to follow the money flows. Politicians will call the AIG loans a success or failure based on ideology, and given the lack of transparancy in government accounting, the ideologues will get away with it by including or excluding flows based on the argument they want to make.

Previously this posting mentioned that in Appendix A “The degree it [i.e., bias] does sneak in is an unavoidable byproduct of their use of a program by program framework for the discussion.” Why is that? It is because the program labels have become synonymous with partisan rhetoric in most minds. Further, deconstructing the programs, as Blinder and Zandi do, shows that most of the programs are such mixed bags of initiatives that the program labels are meaningless other than as hangers for partisan rhetoric. Nevertheless, by deconstructing the programs, Appendix A allows the reader to reassemble them into more meaningful categories.

Interesting thing is that reassigning individual measures shifts some items from a financial measure in TARP into the stimulus category and vies-a-versa. (Retain that idea because it will explain why The Hedged Economist uses a larger number when discussing stimulus than Blinder and Zandi). The reassignment occurs even at the level of identifying measures designed to restore capital flows verses those designed to more directly increase production and employment. So, breaking away from program categories leads to some very different conclusions, but that will be discussed below and in the subsequent posting.

For now, let’s consider some more meaningful categories. This will be broad brush and isn’t intended as a substitute for reading Appendix A. Rather, the reader should view it as a framework for getting the most from the Appendix.

The programs reviewed include loans, guarantees, expenditures, and tax cuts. Each of those can be broken down further. But, even at that high level of categorization, the breakdown is more meaningful than programs.

The loans can be broken down by terms (e.g., market rates, mandated rates, short term, long term, and by collateral provided such as an asset or equity). It is also useful to categorize them by the reader’s assessment of whether they were likely to be paid back (i.e., “probability of default” in street jargon).

Guarantees can be broken down by terms also (e.g., priced put that had to be bought back, priced put that could be bought back, free put options, puts on a specific liability, and whether they were a condition for a loan with the loan classified as above).

Expenditures can be classified by type (e.g., who received the money, who was being subsidized (which isn’t the same as who got the money), what industry or sector benefited, and whether the expenditure was net public expenditure or a substitute for other public expenditures).

Tax cuts can be classified by whose taxes were cut (e.g., business taxes or personal taxes). There’s much more that can be done to categorize the tax cut (e.g., which tax and what groups’ tax), but business and individual is a good start.

Approach the financial measures and fiscal policy during this recession by type of measure rather than program and some really interesting anomalies become apparent. For one, many people will react differently to exactly the same policy response depending on which program it is a part of.

Second, the only legitimate way to identify the actual cost or profit associated with benefits resulting from a measure is to look at the actual measure not what program it was in. It’s the only way to get at the actual cost of the individual measures. Cost can’t be ignored and still hope for a meaningful discussion of cost-benefits. It’s also easier to compare benefits of similar measures if they are identified as such rather than as different programs.

Third, it reveals incredible waste as different policy responses offset each other.
Fourth, it leads to interesting insight into why some responses failed, including the contradictory programs mentioned above.

Fifth, it highlights how effective some approaches were, regardless of which program they were under. Finally, it provides a much more meaningful categorization. A fiscal stimulus is a fiscal stimulus regardless of whether it is in TARP or a “stimulus plan.” A loan guarantee is a loan guarantee. The subsidy value of government measures resulting from loan guarantees are probably more uniform across programs than are tax cuts, expenditures, loans, and guarantees grouped into one program.

One of the conclusions one reaches by abandoning the program by program approach relates to some often quoted investment wisdom. Buffet puts it something like: Be bold when others a scared and scared when others are bold. Bagehot’s version is probably more relevant since he was talking about public policy (although it was for a quasi-private organization at the time). His version is something like: In times of financial panic, lend freely against good collateral at usurious rates and you will always make a profit. There was a fair amount of that approach involved.

Interestingly, some observers object to loans that made a profit, but applaud those that will be written off and vies-a-versa, often with no justification based on differences in benefits. The same inconsistency, perhaps irrationality, surfaces with loan guarantees.

The other conclusion that jumps out when the measures are approached by kind rather than program is that the government did a massive transfer of balance sheet risk. They took the risk associated with ultimate responsibility for a lot of liabilities and potential profit from assets off of private sector balance sheets and moved it to the government’s balance sheet. What’s particularly funny about this is that often it involved off balance sheet items from the private sector and became off balance sheet items for the public sector. It’s telling that the public sector tends to bring the profits from the assets back onto its balance sheet, but leaves the liabilities hanging out there in never-never land.

What’s frightening is that the government also took on ultimate responsibility for some very chancy liabilities without any assets. Some of the loan guarantees are examples of high risk with no potential return while others are extremely low risk.

The big takeaways from this posting are: Read Appendix A of the article even if you don’t read the rest of it. As one reads it, the benefit of re-categorizing the policy responses should be a greater understanding of what was done. Conclusions about the desirability of various policies should be more robust to any bias on the part of commentators in the media. One may not reach the same conclusions as The Hedged Economist, but your conclusions are likely to be more firmly grounded and they will be YOURS.

No comments:

Post a Comment