I have recently seen an explanation for the equity risk premium (which is supposedly higher than some risk models would justify) that seemed reasonable to me. The explanation was that there is an unavoidable observational bias, in that, it is only because we are in an extended period of peace and prosperity that we are systematically measuring ERP, and that kind of period would naturally lend itself to unusually high equity returns. In other words, returns to equities will almost always appear to be slightly inflated, except for those few occasions when they are highly negative - the ERP has a bond-like range of outcomes.
In the end, the universe has an expected range of payouts much like a bond - even a lizard spends months or years hanging out under his bush, munching on a tasty grub now and then, until when he least expects it the hawk's talon hooks him one day while he's basking in the sun.
So, in the end, what has finance gained us? We want to be insulated from risk. Humans, like all social animals, have always depended on familial and community bonds to grasp hands and pull through famines as a group. But, in simpler societies, there were still a tremendous amount of risks, dangers, and famines that were too much to handle, even together.
Modern complex societies allow us to trade and share risk much more effectively. Since most of us aren't in the risk business, we seem to have settled at a social equilibrium where we have bifurcated into risk buyers (business owners, equity investors) and risk sellers (laborers, creditors). Creditors are actually on both sides of this equation - they buy risk from laborers and sell it to equity holders.
Labor and creditors share that same payout - they each produce a premium for the risk buyers (the equity holders), and in exchange, they receive very standardized, predictable outcomes....., until the day that default or layoffs come. I wonder, in the end, how healthy this is. Morally, it allows us to pretend that the tendencies of the universe are suspended, that abundance is secure and predictable. But, we've really just paid the equity holders a premium to take all the noisy, manageable risk from the fat part of the range of outcomes.
I think this might lead to biased moral reactions to risk. Equity holders have explicitly acquiesced to accept a certain kind of risk, so we see their success or failure as acceptable. But, since laborers have sold most of their small risk away, we see the imposition of risk on them as less acceptable. There was never any evidence that anyone could have accepted the long tail risk, but our daily experience in the world of normal outcomes, we compartmentalize people among the risk bearing and the non-risk bearing. We hiss at equity holders who layoff workers in order to move a factory to a less expensive location, but we don't take the same offense if the factory stays open but goes bankrupt, leaving the jobs and bondholders intact, but the equity holders with nothing. These are both the result of unfortunate long-tail outcomes. Is the moral double standard justified?
This might explain an additional ingredient into the ERP, because this moral framing causes us to create social policy countering these moral reactions. Tax policies and other regulations tend to favor debt over equity and labor over capital, depositors over banks, etc. But, even these policies cannot overcome the broader universal reality of long tail risk. So, they simply create more demand for the kind of seemingly risk-free trade offs that push more investment out of broad risk-exposure and into this binary payoff world. This payoff structure, and the excessive dependence on it in an economy, creates more danger of systemic risk, because there are fewer participants willing to take on economic risks explicitly. The system becomes less robust. And, when systemic crisis hits, our moral predisposition leads us to politically forgive the long tail risk of the risk averse. Bond holders are bailed out, while equity is wiped out, labor protections and insurance are expanded. These are more or less predictable reactions which increase the de facto long term cost of risk, and would, theoretically, increase the ERP.
An extreme example of this phenomenon is in banking, where the federal government guarantees deposits. Deposits are a form of debt, from the perspective of the bank. Additionally, a form of financing called a repo, which is essentially a loan that is given favorable legal treatment, has also become popular. So, banks have these tremendous sources of debt financing with very favorable legal treatment. This has the effect of pushing up the debt ratio that banks use. The government counters this effect with capital requirements, so in order to compete, the banks have to leverage up to the maximum allowed amount of debt. Normally, as leverage rises, debt-holders are less protected from downside risks, so they demand a higher premium, in the form of higher interest rates, and a balance is reached between the levels of debt and equity. In non-banking industries, typical leverage levels depend on the factors in a given industry, but they are always naturally much lower than in banking. Since accounting and government policy insulate debtors (including repo buyers and depositors) from financial risk, the debt on bank balance sheets increases with no natural impediment. As long as these policies remain in place, if banks didn't also have capital controls, they would be essentially fully leveraged 99.9999%, with the small equity holder taking gains while they came, and leaving the FDIC with the remains when failure encroached on the organization. In earlier banking regimes without these regulations, bank leverage was much lower.
When crises hit, the equity holders are wiped out, and generally we undertake additional political policies to salvage the bondholders and depositors. Equity in banks should call for an extremely high risk premium in this context.