At first glance, this post may seem rather pessimistic. When people ask me what’s going on in the economy, they don’t want to say that macro variables are difficult to predict. Still, I see this as a promising post. Writing it actually made me more optimistic about the forecast.
Before explaining my theory, let me review two similar but clearly different theories, Efficient Market Hypothesis (EMH) and Lucas Criticism:
EMH says asset prices are difficult to predict, as current asset prices reflect the expected impact of already publicly available data. Thus knowing that Tesla car sales are growing rapidly and governments are emphasizing green energy does not help me to predict the rate of return from investing in Tesla stocks. This information has already determined the price in the market.
Lucas Critics says that when policymakers try to take advantage of the historical relationship between policy materials and a policy goal variable, the relationship will change and become unstable. Thus if you notice that there is a positive relationship between the level of money supply and employment under a gold standard and then artificially increase the supply of money to create jobs, there will be a tendency for the relationship to break down. Workers will start demanding higher wages in anticipation of higher inflation in the future.
None of these theories interfere with my or anyone else’s ability to predict macro variables. I am not a policy maker, and thus Lucas Critic does not apply to me. And EMH does not rule out the possibility of being able to predict rising inflation or recession in 2023, as those forecasts may already be linked to asset prices. However, these two well-known theories are somewhat similar to the hypothesis I am going to propose, based on three hypotheses:
1. Represents most of what we are told to predict Policy failure. Not all predictions; It is certainly possible to predict a healthy economy. But the predictions that people value the most are policy failures, such as rising inflation or the next deep recession.
2. We often make predictions by looking at past patterns in the data. We say, “The last time X happened, the experience of economics was Y.” Importantly, the “X” is almost always universal information.
3. Policymakers generally try to prevent policy failures and rely on public information.
Every time a large plane crashes, investigators recover the black box and try to find out the cause. If a component fails, they may ask the airlines to replace that component with something more reliable. If this is a pilot error, they can tell pilots what went wrong and how to respond to the situation more effectively next time. As a result, it is really difficult to predict what will be the cause of the next major plane crash.
Most macro forecasts contain slightly more than what economists observe: “In the past, I have noticed that macro shock X was often followed by policy failure Y.” If policymakers do not learn from their mistakes, it will be an effective way to forecast macroeconomics. But policymakers learn from their mistakes. They don’t learn as much as I want to learn quickly and effectively, but they learn. And that learning (combined with subsequent adjustments in policymaking) makes macro forecasting much more difficult than otherwise. Indeed, this point holds Even if the policy makers do not get the wrong educationবলShow over-responses where in the past they reacted less. Any adjustment in policy based on education makes prediction much more difficult than otherwise.
In my view (and the optimistic part of the post here), it gives us two useful ways to predict.
1. Not all bad results reflect future policy mistakes. Some bad results may be less than bad considering the previous policy mistakes that have already happened. For example, when there was great inflation (due to additional monetary stimulus), the Fed briefly adopted a tight monetary policy in late 1966 and early 1967, which slowed NGDP growth to about 5%. Fearing a recession, they moved away from that policy and in the next 14 years NGDP growth increased to an average of more than 10%. In retrospect, their fiscal restraint (say 5% NGDP growth) should have continued even though it turned into a mild recession in 1967. The alternative (great inflation) was even worse.
Today, the Fed’s NGDP growth rate needs to slow to more than 4%, perhaps a little less. Doing so increases the risk of recession, but it is still worth doing. This fact allows many people today to confidently predict a recession, where a recession is much harder to predict when the economy is in balance with low inflation and high employment, and any recession will represent a policy error. Thus bad results can be predicted when they represent the best policy – less bad in dealing with an already bad situation.
2. Another way to predict bad outcomes is to look for evidence that policymakers have not learned the right lessons. In 2020 and 2021, Bob Hetzel J. Powell looked at the rhetoric coming out of the Fed and noticed the annoying parallels with the policy that created the great inflation. The Fed learned some useful lessons from the mistakes made during the 2007-09 Great Depression, but shows additional reactions because it ignored lessons from the 1960s and 1970s.
In short, any attempt to predict bad macro results involves a combination of two types of analysis. First, to make sure when bad results are almost inevitable, because they represent less of the bad (often due to previous policy mistakes.) Second, to try to understand what kind of mistakes a particular set of policymakers can make.
But we should also not ignore the pessimistic side of this analysis. History cannot play almost the same way as policy makers always learn from past mistakes, even where they learn wrong lessons or are part of a true story. We try to predict bad times for the economy, Jay Powell is trying to make us fail. And he has very powerful tools in his hands.
No amount of progress in the science of macroeconomics can solve this problem, as it is essentially an arms race between the forecaster and the Fed.