One small footnote to the Brexit controversy: It has invigorated my interest in Lord Mervyn King’s concept of radical uncertainty. As the former governor
of the Bank of England laid out so eloquently at PIMCO’s investment forum in May and in his recent
book The End of Alchemy, “radical uncertainty” refers to uncertainty so profound that it is impossible to describe the future in terms of a
knowable and exhaustive range of outcomes to which we can attach probabilities. I find this concept particularly useful to describe the challenges we all
face in a world of rising polarization, populism and politicization.
The King is dead, long live the King!
Lord King argues that radical uncertainty is pervasive and that the inability to conceive of what the future may hold means that the probabilistic models
that economists, central banks and investors use for forecasting are doomed to fail. To quote from King’s book (p. 304):
“In a world of radical uncertainty there is no way of identifying the probabilities of future events and no set of equations that describes people’s
attempt to cope with, rather than optimize against, that uncertainty. … In the latter world, the economic relationships between money, income, saving
and interest rates are unpredictable, although they are the outcome of attempts by rational people to cope with an uncertain world.”
This is truly remarkable stuff from an ex-central banker who pioneered and perfected inflation targeting with its heavy reliance on economic models and
forecasts, and who, in a former life, was (to quote from Paul Krugman’s recent review of King’s book) “a
card-carrying mainstream economist.” Obviously, the 2008 financial crisis and the unusual New Normal macro environment ever since have made King reconsider
and eventually break with the conventional approach and come up with a new one – a rare and laudable occurrence among economists and central bankers!
Re-enter the Knight
Of course, as King notes himself, the concept of radical uncertainty isn’t really new. Economists have usually referred to it as “Knightian uncertainty”
ever since University of Chicago professor Frank H. Knight in his
1921 book Risk, Uncertainty and Profit distinguished between “risk,” which can be quantified by attaching probabilities based on experience and/or
statistical analysis, and “uncertainty,” which is essentially unmeasurable and represents the unknowable unknowns.
However, even though the concept of Knightian uncertainty has been around in economic thinking and John Maynard Keynes also discussed it (without giving
credit to Knight as far as I can tell) in Chapter 12 of his 1936
The General Theory of Employment, Interest and Money
it was largely forgotten or ignored by the following generations of economists, who became more interested in formalizing their discipline and building
seemingly precise models of the economy based on the idea that we can attach probabilities to future events and outcomes based on observations from the
past. And sadly, it took the global financial crisis of 2008 to lead Lord Mervyn King and others to rediscover the idea that many future events are simply
impossible to conceive of today and to capture in economic models.
Stuff happens, and more often than you think
Of course, it is debatable whether one should go as far as King and declare radical uncertainty as so pervasive that all forecasting is futile. My hunch is
that more often than not, and particularly over shorter time horizons, we can assume that we are in a period of overall stability, where insights into the
way past events unfolded can assist in forecasting the future. However, we always have to be aware that the real world is far from stable or stationary and
that, to quote King once more, “stuff happens” – think Lehman, Greece, Brexit and Trump, which can or will lead to regime shifts that render the old
empirical relationships obsolete.
So what does the presence of radical uncertainty imply for analysts, central bankers and investors? Let me offer three thoughts:
Think the unthinkable, or at least try to
First, economic and market forecasts based on statistical/econometric models still make sense, but keep in mind that they always rest on the
often-forgotten assumption of structural stability. Regime shifts and structural breaks (“stuff happens”) are more frequent than most people tend to
believe. I’ve been doing applied economics for more than 30 years now and there have been more “structural breaks” than I have fingers on my two hands
(yes, I’m a two-handed economist!). That’s why, in addition to working with models, it is important to think in terms of scenarios, engage in “what if”
analysis and, importantly, force yourself to “think the unthinkable” – the trends or shocks that statistical analysis would tell you are highly unlikely
because there is no precedent.
Forget forward guidance
Second, if the future is radically uncertain, then the modern central bank practice of giving markets forward guidance may be, well, misguided. Yes,
central bankers always emphasize that guiding doesn’t mean promising and that, e.g., the Federal Reserve’s dot plot of an appropriate policy rate path
is no more than a conditional forecast. Yet, if these forecasts turn out to be wrong most of the time because “stuff happens,” what is the value of making
them in the first place?
Perhaps, it would be better to acknowledge the existence of radical uncertainty and stop making policy forecasts that will inevitably be proved wrong most
of the time. So far, the Fed has only made a small step in this direction by calling itself “data dependent.” Yet, by publishing forecasts for economic
variables and its own interest rate path, it still creates the illusion that the future is highly predictable. As I see it, true data and regime dependence
is not compatible with explicit forward guidance. And, as an aside, doing away with the dot plot and the guessing about how it will evolve and who is who
in the dot plot would free up many analysts’ time to think more about the unthinkable.
Populism promotes even more radical uncertainty
Third, radical uncertainty may be even more acute today than it has been over the past several decades because we seem to be entering an era where politics
again dominates economics in shaping market outcomes. “Political economy” variables usually don’t enter economic forecasting models. Yet, as King rightly
points out, it has often been big surprises on the political front that drive major developments in the world economy. They represent “not the random
shocks of the forecasters’ models but the realization of radical uncertainty,” as King phrases it. Brexit, Trump and the rise of populism – what are the
consequences, and what is next? We can all make educated guesses (see for example my take on how Brexit could lead to a stagflationary outcome). But only time will tell.
Not so incidentally (as King spoke at our Secular Forum in which we looked out three to five years), the concept of radical uncertainty also played into
our forum conclusions, where we characterized the longer-term global outlook as “
Stable But Not Secure
.” Yes, the global economy and financial markets appear relatively stable, and that stability may well last a while longer, perhaps even beyond our
cyclical (six to 12 month) horizon. However, a deceptive, treacherous stability it is, as risks are lurking and building in the background – ever more
elevated asset prices relative to historical norms, mounting private and public debt, diminishing returns to monetary policy and the rise of populism. And
these are just those risks we are able to identify. All of this, and the uncomfortable but uncanny presence of radical uncertainty, leads to a very simple
conclusion for investors: Let capital preservation be thy King!