The sorry state of banking research

Economic academic research can be curious. In particular since the financial crisis, academics have focused on proving that free markets were inherently unstable and that government intervention was required to stabilise the economy.

While George Selgin incinerates a recent paper on Canadian private currency, I found three other recent papers that try too hard to convince us that markets aren’t perfect.

The first one, titled Short-termism Spillovers from the Financial Industry, attempts to demonstrate that large listed banks are subject to short-termism in order to meet quarterly earnings figures, and this that short-termism affects their behaviour towards their clients and in turn borrowers’ long-term investment policies. They conclude that short-termism is not optimal from an economic efficiency point of view.

In their words:

First, we find that lenders facing incentives to meet quarterly earnings benchmarks are more likely to extract material benefits from borrowers. Second, lenders with short-termism incentives push relatively high-quality borrowers into material covenant violations because these are precisely the borrowers from whom rents can be extracted. Because unhealthy borrowers are already selected for material covenant violations by lenders both with and without short-termism incentives, only relatively healthy borrowers are left to be targeted by incremental attention. Third, affected borrowers are more likely to reduce capital investment and research and development (R&D) expenditures. Given the selection of higher quality borrowers, it is particularly likely that these real investment effects on borrowers are value-destroying. Finally, we find that the market reaction to announcements of material covenant violations is 88 basis points lower among borrowers whose lenders face short-termism incentives, which suggests that the incremental attention from lenders with short-termism incentives does not improve shareholder value.

While they fall short of recommending government or regulatory intervention to maximise value-enhancing investments, the implication of their paper is clear: free markets do not optimally allocate capital. But don’t take their word for it. This paper is highly problematic for a number of reasons, outlined below.

First, they use equity analysts’ consensus earnings per share forecasts as a benchmark for short-termism, despite the highly inaccurate nature of those estimates. Indeed, those are constantly revised ; and banks are fundamentally very difficult to model due to the opacity of their balance sheet. As a result, analysts’ estimates are often wide off the mark and do not represent a reliable indicator. Sadly, the whole logic of this paper rests on this single benchmark.

Second, this paper makes rather strange assumptions about the utility of covenants in loan documentations. Covenants are usually agreed upon during the negotiations of the lending facility in order to protect the lenders by preventing the borrowers from fundamentally altering the nature of its balance sheet or of its business model. A breach of covenant is a contractual breach that is considered a serious event by the lenders as it implies a decline in asset quality. Yet this paper seems to argue that enforcing covenant is a bad thing, which ends up negatively impacting the borrower’s ability to grow in the long run.

They go as far as qualifying covenant enforcement as ‘extracting rents’ from borrowers. This is incredible: covenants are rules that are in place for a reason. Not to enforce them on a regular basis would undermine the effectiveness of those rules altogether and probably lead to much worse outcomes. Moreover, researchers qualify some of those covenant-breaching borrowers as ‘high-quality’ and ‘financially healthy’. I can assure you that, in the real world, covenant-breaching customers are anything but ‘high-quality’ and are usually flagged as ‘risky’ by bankers.

Third, even assuming their logic and methodology are correct, the effects they find is small: they calculate that borrowers affected by enforced covenant breaches are only 2.4% more likely to cut R&D spending and 4.9% more likely to cut capital expenditure. Borrowers are also only 1.4% more likely to switch lenders for their next loan and financial market reactions are marginal (88bp). Talk about a storm in a teacup.

But more importantly, my main concern is that the authors of this paper never ever benchmark their results. Or, more accurately, they benchmark the results they obtained against a hypothetical ‘social optimal’. As such, they fall in the Nirvana fallacy trap that Selgin also refers to his post: free markets are not perfect but no amount of government intervention could fix those admittedly minor shortcomings.

The second one is titled Macroeconomics of bank capital and liquidity regulations and studies the welfare effects of banking regulations. Or rather, it ‘models’ this welfare under very specific assumptions. So specific actually, that I dismissed the paper straight away.

In my view this paper exemplifies a lot that is wrong with today’s current economic research: it is based on a highly theoretical mathematical model with imbed assumptions and limitations that result in outcomes that do not nearly reflect the real world. Yet, those economists still managed to conclude that “capital and liquidity regulations generally mutually reinforce each other”, and that “the optimal regulatory mix consists of relatively high capital and liquidity requirements” (which they define as a very high 17.3% leverage ratio, more than ten percentage points above that of most banks today). They evidently conclude that their analysis provides broad support for Basel III’s regulatory framework, consequently seen as welfare-enhancing even though it doesn’t go as far as those economists would like.

Well… the one huge issue with this paper rests on this particular assumption underlying the trade-off faced by regulators in their mathematical model:

On the one hand, banking regulation may reduce the supply of credit to the economy. On the other hand, it improves credit quality and allocative efficiency. Accordingly, regulation tends to result in less, but more productive lending.

This is my reaction to this sort of nonsense:

There is not a single glimpse of reality in believing that regulation is more effective than free markets at allocating capital in the economy. If anything, as I highlighted so many times before, regulation has diverted the allocation of credit from productive uses (i.e. commercial and industrial loans) towards unproductive ones (i.e. real estate), which has been economically damaging and one of the reasons behind the financial crisis.

As a result, this paper includes some of its own conclusions in its assumptions, leading to circular reasoning: banking regulation improves allocative efficiency, therefore we need banking regulation.

Finally, Bank Capital and Dividend Externalities highlights that banks fail to ‘internalise’ the effects of dividend payments and capital policy on the stability of the wider financial system. The researchers theorise that banks increasing their dividends harm the claims that its own bank creditors have on its balance sheet in a bankruptcy scenario, thereby weakening the financial strength of the whole network. In such a system, they state that bank capital becomes a ‘public good’. The logical conclusion to this lack of systemic coordination is obviously government intervention: regulators should put dividend restriction measures in place when necessary.

But, here again, this paper suffers from major design flaws:

– Bank creditors are the only ones considered, despite the fact that banks have a multitude of creditors, including depositors. If dividend payments harm bank and other money market creditors, they also surely impact depositors and bondholders.

– They assume that dividend payments decrease the value of the bank by lowering its probability of survival. They reject the signalling theory of dividend payments despite admitting it had some backing in the literature: reducing dividend is a negative signal about the financial health of the institution.

– Their empirical evidence is limited to a couple of data points taken during the latest financial crisis: a couple of banks increased dividends before collapsing. They do not take into account the fact that those followed Bear Stearns’ bail-out by the Fed, which sent a specific TBTF signal to both the market and bankers themselves.

– As often in the banking literature, they make a big deal of interconnectedness, yet seem to forget that all industries have interconnected members. The decisions made by a large automaker also affect its whole supply chain and their employees. I have yet to see tens, if not hundreds, of research papers arguing that automakers fail to ‘internalise’ the impacts of their decisions, which are not always ‘socially optimal’.

– More damning, they prove their theory by modelling a financial system that includes just two banks, consequently suppressing any opportunity for exposure diversification. This is completely unrealistic. Banks have tens of other banking counterparties and already factor in their counterparty assessment the possibility that capital policy might change. But thanks to the diversification effect, those changes usually only affect them at the margin. A two-bank model does not capture this critical point. To be fair, those economists do admit that their model is a simplified one. Yet this admittedly weak theoretical basis does not seem to make them think twice before making policy recommendations.

– Finally, this paper falls into the same Nirvana fallacy trap as the first one reviewed above: they do not make a convincing case that an effective alternative exists, and assume away Public Choice issues by relying on the actions of omniscient regulators.

This is the sorry state of affairs in current economic research. By focusing on highly theoretical mathematical models based on very limiting – if not totally unrealistic – assumptions, damaging policy recommendations are outlined and subsequently serve as justifications for regulators’ actions. Clearly, nothing much has changed since the days of Diamond and Dybvig’s flawed model (see White, The Theory of Financial Institutions, 1999). As I recently pointed out in the case of macroprudential research, this is a reminder that it is critical to read research papers’ body (and not only their abstracts and conclusions). Sadly few people seem bothered do so.