Many books and articles have been written to describe the mishaps that led to the Great Recession of 2008, but few have done so while putting the role of accounting standards and auditing in context. I reported back in January 2010 that Paul Krugman at least was cognizant enough of accounting to acknowledge that we do not presently have honest financial reporting standards that let investors make good decisions and that "… forces management to behave responsibly." But, other than changing the name on the sausage packaging from GAAP to IFRS, detailed suggestions for achieving fundamental change have been exceeding rare.
NC State accounting professor Paul Williams also drew inspiration from Krugman's column and the financial crisis for a brief paper entitled The Wrong Metaphor, which I recently heard him present to the Public Interest Section of the American Accounting Association. Much as I have (but certainly with some important differences that you should read for yourself), Williams has come to the conclusion that a paradigm shift is needed for accounting to re-assume its proper role in society. I wish we could have much more from Williams reagrding what he thinks should be changed and how to make change happen.
A recent paper by S.P. Kothari and Rebecca Lester of MIT, The Role of Accounting in the Financial Crisis: Lessons for the Future, is very strong in its description of flawed accounting and its potential role in the past crisis. These are some of the basic messages:
- The accounting rules provided originators of subprime mortgages (SFAS 140) the opportunity to record economically specious gains on securitization transactions.
- The accounting rules also permitted regulated financial institutions that invested in financial instruments created out of subprime mortgages (MBOs and CDOs) to delay recognition of likely losses; which, in turn, distorted measures of capital adequacy.
- Estimation errors, most significantly estimates of fair value, committed by issuers of financial statements and their auditors exacerbated the accounting misstatements inherent in the accounting rules.
Although, like Williams, the Kothari/Lester did not provide much in the way of "lessons for the future," I would still highly recommend it. Accounting students could, and should, read it if only for its lucid examples and explanations of the arcane and flawed accounting rules that are applicable to securitizations.
The purpose of this post is to distill the analysis of Kothari/Lester so as to further discuss "lessons for the future." Before I do that, however, I want to tell you about an email correspondence I had with S.P. Kothari after I read their paper. Since I was preparing my own remarks for academics on the topic of IFRS adoption and the future of U.S. accounting, I asked him whether he thought it more or less likely that promulgation of more appropriate fair value or securitization standards would occur if the SEC continues to pursue convergence with IFRS. This was the response:
"…I think convergence will harm promulgation of more appropriate standards because a converged standard setting body would become a monopoly that would be pulled in all different directions by its constituencies."
Clearly, whatever the "lessons for the future" may be, adoption of IFRS is not one that he had in mind.
Securitization Accounting as Grand Theft Auto (not the video game)?
The process of securitization reminds me of large-scale car theft ring.* To obscure the link to the illegal 'acquisition' of cars, much like subprime mortgage loans, the cars are almost immediately chopped up into parts and re-sold. As a business model, stealing cars works only because the cost of obtaining a car is low; but let's ignore that for the moment and assume that someone actually establishes a legit buy/chop auto parts business – in other words, cars are bought for full price and chopped up for the sale of parts. Here is a representative sent of journal entries:
Day One – buy a car:
Dr. Car (at cost) 30,000
Cr. Cash 30,000
Days Subsequent – chop up the car and sell its parts:
Dr. Cash (from sale of parts) 25,000
Dr. Engine (at estimated fair value) 7,000
Cr. Gain on sale of parts 2,000
Cr. Car** 30,000
I readily acknowledge that second entry does not comport with current GAAP. I am using to illustrate securitization accounting in the context of a highly simplified example. Having stated that, however, I don't see any reason (although others might) that the accounting for a chopped car should be different than the accounting for a chopped subprime mortgage. I can't abide mixed-attribute accounting systems, and believe that college-educated professionals should not be adding apples and oranges, much less elevating the error with misleading names like "statement of financial position."
If you can set aside any cavils you might have with respect to my violation of current GAAP, please answer this question for me: Would you invest in a business whose financial statements were based on journal entries like these? They sure do look legitimate, but common economic sense would dictate that the second journal entry must overstate the real profit on the resale transactions. First, as Kothari and Lester observe in the securitization context, "how is it possible that the sale of an asset originally recorded at one value could generate income and higher asset balances upon merely transferring this asset to another entity?" I suppose one could say that butchering a chicken should add value such that the sum of the parts is greater than the whole chicken. But that's because there's only one way to one way to obtain chicken parts – first, you gotta grow a chicken. However, one doesn't have to manufacture a car just to obtain a spark plug; or to originate a subprime mortgage loan just to create contracts containing a particular pattern of future cash flows.
The second reason indicating that the second journal entry is misleading is that the flawed business model would not be able to earn an economic profit over a full cash-out-cash-in cycle (i.e., buy car Þ chop Þ sell Þ reimburse for defects under warranty). If the engine could actually be sold for $7,000 (and generating an overall profit of at least $2,000), then I'll be a monkey's uncle!
The Lessons for the Future
In order to determine whether a better accounting is feasible, we need to be able to identify the sources of overstated income. With respect to both the car parts example and securitizations, overstated income could be the result of any of the following:
- Kothari/Lester do not refer to this explicitly, but the most subtle and important aspect of the problem, in my view, is the unavoidable change in the 'unit of account' from the original asset (e.g., car, subprime loan) to the retained portions (e.g., engine; or in the case of securitizations, a servicing asset, and the right to residual cash flows). For example, erroneous accounting will occur because the values of the components of the car will not exactly sum to the pre-chopped value of the car itself; it doesn't matter how one chooses to measure value.
- The appropriateness of the basis of measurement for the retained assets: such as historic cost, fair value or replacement cost.
- Given a basis of measurement, biased measurement of the assets retained (which auditors fail to correct). Kothari/Lester surmise that this is the most significant source of overstated income.
- Unrecognized or understated liabilities (e.g., a warranty obligation on the car parts, or a written put option for underperforming loans).
As to mitigating these sources, I have provided my two cents on #3, the biased estimate problem in a number of posts, the latest of which is here. Regulators have tried over and over again to make incremental improvements to auditing standards, with no perceived help in increasing the public's confidence that auditors can play a significant role in averting a large-scale financial crisis. It is time to abandon any notion that: management can provide unbiased estimates of value; and/or that auditors are intellectually capable or even inclined to provide a independent and knowledgeable assessments of management's estimates.
Therefore, when financial statements require estimates of current values and uncertain future events, these estimates should be produced by third parties – i.e., neither management, nor the auditor. The auditor's role should be limited to:
- Verification of the components of financial statements that can be verified.
- Verification of the inputs to current valuations that can be objectively measured, and hence, verified.
- Verifying that third party experts charged with measuring current values and expectations of future events have performed the procedures they said they would perform.
As to #2, above, the conventional wisdom is that historic cost accounting is fact-based, objective, understandable, and even verifiable. We have learned the hard way that this is a fiction; however historic cost accounting has been practiced. To get to a point where financial reporting can be freed from perverse incentives for management to produce biased estimates, standard setters must dedicate themselves to the establishment of accounting standards that yield financial statements capable of being audited and understood by auditors, investors, credit rating agencies and the like. Securitization/loan accounting standards clearly indicate that historic cost "principles" when applied to financial instruments create nothing but a big mess.
Saving the worst for last, the unit of account issue (#1, above) may defy a principles-based solution, but I still want to take a stab at it. Referring to the sample journal entries, the higher the values assigned to the assets retained, the higher will be the reported gain. This suggests, once again, replacement cost is the most appropriate basis of accounting, for it would require measurement based on the least costly way to replace the asset subject to being measured. For the car example, the business model indicates that the most economical way for the entity to acquire an engine would be to buy a car, chop it up, and sell all of its parts except for the engine. Thus, the marginal cost of an engine might be essentially zero – and the result would essentially eliminate the gain recognized in the second journal entry. Indeed, the loss recognized in the second journal entry by this logic would be the telltale sign of a flawed business model.
The same logic used to measure the replacement cost of the engine could easily apply to securitization assets such as servicing rights and residual interests, and leads to an analogous question: Did the subprime loan origination/securitization business model only work, or appear to work, because of favorable accounting treatment? These are questions worthy of further research, but if the accounting for securitizations had been more reasonable, we would already have reasonable answers.
*Kothar/Lester address both the accounting for securitizations by the originators of subprime loans and the subsequent accounting by investors in the financial instruments created by securitization. I am only looking at the former in this post.
**I am assuming that costs of chopping the car are negligible or are expensed in a journal entry not displayed. Capitalization of these destruction costs would affect the form of the journal entries, but it would not affect my conclusions.