I have the house to myself today, and after polishing off some off UCC’s and Kirkland’s finest, I decided to look over a few spreadsheets of statistics, since that’s certainly my ideal Saturday morning. Actually, it’s pretty darn good, but I digress…
The Zero Hedge post I mentioned previously was a good general chat about how policy makers and really, anyone tends to latch onto easily understood indices, models, market indicators, etc. In particular, the attachment to controlling inflation is endemic among central banks and politicians hooting and hollering at rallies. It’s really just a phenomenal example of a heuristic for those who don’t know enough to know that they cannot possibly grasp all the elements that affect an economy’s long run trajectory.
While there’s assuredly a case for controlling extreme inflation, if we engage a bit of epistemic humility and think objectively about the scale of national economies, it’s ridiculous to think that anyone could in definite terms say a single metric if kept at desirable levels represents a healthy economy. Daiichi ni, how is one to determine what that desirable level actually is? By using models based on mountains of historical data of course!. Let’s see that famous “Taylor Rule”…
Run some algorithms over heaps of past data and ideal values for those two coefficients should pop out, and with an approximation of potential GDP and some (arbitrary) inflation rate target in hand, to our delight out comes the ideal nominal federal funds rate… but wait, aren’t most complex systems characterized by the fact that they change over time? So, wouldn’t that mean the present is different from the past? If so, using parameters which described the system when it was in an entirely different state for predictive measures doesn’t leave me with much confidence.
Yet, thanks to a false sense of confidence instilled by (as an associate of mine puts its) the “we have all the tools we need to solve our problems” narrative which acts as a backdrop to the countless economic summits, the people at the top continue to dictate what is and what isn’t an end-all-be-all economic indicator.
The Fed recently set a 2% inflation target, which sparked some criticisms of the Bank of Japan’s current “lackluster” policies1… and so the cycle continues.
Since the late 1980s, the Japanese Consumer Price Index has indeed been pretty stable, but I’m going to go out on a limb here and say that some other indicators weren’t quite so stable. In fact, one might be able to say that it is precisely because of excess resources being poured into micromanaging consumer goods prices that other sectors of the economy see volatility (thought certainly some are just inherently volatile unless controlled).
First a look at CPI and the land prices in the top 6 urban areas (it’s essentially the same trend even if looked at nationwide), and below that, a comparison of the year on year rate of change for the two indices.
How about government bond issues? Left vertical axis in billions of yen:
It looks like prices sure have stayed good and calm recently – I wonder where the capital to fund the constant interventions came from? Oh…
This next one is perhaps gratuitous, but it isn’t an exaggeration in the least to say that even if one doesn’t own a single financial instrument (besides bank deposits, which of course is the most ubiquitous of them all), one’s savings are tied to a degree to the movements of capital markets, thus let’s see how the Nikkei shuffled around the past couple decades:
Again, while many financial instruments are just derivatives of totally separate underlying assets, since a great deal of the real wealth people have stored in banks can potentially be exposed to such instruments via the bank’s proprietary trading business, by the same logic which drives policymakers to strong-arm interest rates to x%, we could then justify doing the same thing to the stock market, or the swaps market, or the CDS market, and so forth. There is no limit to what we could potentially label as an economically significant indicator worth our attention and worth mechanically tweaking to oblivion. Yet, in looking at the results of such specifically targeted policies, it is clear that there are forces at work here which are too complex for us to grasp fully, thus the consistent failures!
While technically speaking, were one to have some computer which held all pricing information about all sectors of all goods and financial markets, updated by the second, and which then fed those changes into models which also immediately adjusted for the revision, addition, and removal of exogenous variables, it might be feasible to reliably predict what a particular policy’s impact would be at a given point in time assuming it went into effect instantaneously. However, policies, and even the information about said policies which travels far faster, is not instantaneously delivered to all economic agents, nor is it interpreted in the same way, thus making predictions about a policy whose impacts suffer from lag essentially impossible.
We may have all the tools we need to solve our problems, or we may not, but the biggest danger here lies in the misplaced belief that we know how to use said tools. Recognizing the limitations we face in our ability to predict things about complex social systems is the pivotal first step towards generating models which have some genuine use!
–
1. FYI: the politician was former cabinet secretary Nobutaka Machimura (町村信孝).