When I was a kid, I was told “mastering mathematics, physics and chemistry will prepare you for a bright future”.
When I got into university, everyone seemed to be learning two languages, one is English and the other was programming. It was then said “English and programming are the must-to-have skills in the 21st century”.
After many years learning, I found myself still confused about many things in everyday life, which cannot be explained by any of those I have been taught in class.
Why are there rich and poor people?
Why do some people believe in Christianity and others in Buddhism?
Is history always repeating itself?
Now, as a parent, I have to answer all sorts of “weird” questions, from the biggest planet in the universe to the fastest car that was ever built in history. After answering more and more questions from kids, I found myself less and less confident in my existing knowledge base.
Furthermore, I start to worry about those questions that have confused me, and the answer to which seems not to be found in any textbooks. There is no doubt that the next generation will face the same problem, and they will probably experience the same confusion as me: “why what I learnt cannot explain what I saw in everyday life?”
To me, it is a shame that I haven’t had the chance to learn economics and politics at an early age. In my opinion, it is more important for ordinary people to get to know the basics of economics and politics than any other knowledge, and if you don’t care about them, they will care about you.
You probably don’t need to know how many planets are in our solar system, but you’d better figure out how to get a better tax refund. You probably don’t need to know how many steps are in the process of photosynthesis in plants, but you’d better understand which political party’s policy can benefit you more.
As an ordinary person, you don’t have to be an economist or political expert, but at least you should try to know the facts, use your judgement skill to analyse, criticize and then reach your own conclusion.
Knowing the facts, for many people, it is easier to say than to do. It is even harder in such an era when everyone can be a source of “fact” by publishing in social media at the speed of the internet.
How to check the facts?
How to use the facts to formulate and support your opinion?
How to build your argument upon facts to convince others?
I think there must be some matured and systematic way to learn the basic skill to build up rational thinking. However, at the moment, I can probably start training myself in my own way, which is more like doing a scientific experiment following the scientific method.
For any economic or political phenomenon observed in everyday life, we need to ask or define a question for it. To answer the question, we can first try to explain it with hypotheses based on any existing theories, common opinions or new ideas.
Then, try to make a prediction or induction from the hypothesis, and search and check as many facts as possible. Next step, use the facts to test the prediction, analyse the result and trace down the reason why the facts confirm or contradict the hypothesis.
Finally, based on the analysis, reach the conclusion to either use the hypothesis for more tests and revise it if needed, or discard the old hypothesis and propose a new one.
In particular, we can try to apply the above steps to a practical problem to see how it works.
Phenomenon: When recession comes, unemployment rate goes high.
Question: Why do people lose their jobs?
Hypothesis: People are lazy: they choose to not work due to individual problems, such as illness, vices, lack of skills or just don’t want to.
Prediction: If some people choose not to work, we should see the unemployment rate at a relatively constant level, or random fluctuation because the individual choice cannot be coordinated or synchronized.
Fact 1: During the Great Depression in the 1930’s, the unemployment rate went up quickly in multiple industrial countries in the same period of time, as high as 30% in some countries.
Analysis 1: In the case of depression, the wide range unemployment is definitely not an individual problem, but a macro economic problem that needs to be further investigated.
Fact 2: We do find some countries with constant levels of unemployment rate for many years. For example, China, India and Vietnam have been maintaining unemployment rates at lower levels than world average, whereas South Africa, Iran and Libya are at 2 to 5 times as high as world average.
Analysis 2: Those constant high and low levels are probably not only due to individual choice, but also reflect the country’s economic structure and particular political and social policies.
Fact 3: The unemployment rate of G7 and OECD countries are showing cyclic and synchronized patterns during the last 3 decades.
Analysis 3: The unemployment rate in traditional industrialised countries have shown similar trends during specific periods, which indicates that it is a much broader problem across countries rather than an individual problem.
Conclusion: Unemployment is not an individual’s choice, but a social problem with wider impacts across multiple countries, and its cause needs much further investigation at a more granular level.
Looking back at the history of economic research, it is obvious that the quantitative and scientific method has been adopted by mainstream economists. Having said that, it is still a big challenge for economists to test their theories by repeatable experiments. In other words, to conduct experiments in the real economy is too costly to be feasible.
However, with the advance of computational capacity, the abundance of economic data collected at an ever-increasing granularity, and newly developed theories focusing more on individuals behaviors, we should see the progression of economic research towards more behavior-driven and data-oriented methodology.
Ideally, any new economic theories could be tested in a simulated economy, where almost every individual’s economic behavior and decisions can be mocked based on data from real life, and the chain-reactions in economy can be simulated, with the required economic indicator being monitored and collected in large scale with high accuracy.
The economists can conduct such an “experiment” in short turnaround without any material impact to the real economy, and the result will be used to analyse and verify the theory multiple times before it turns into the real economic policy.
There are no new things under the sun. Apparently, some economists have been on the way.
A research team from Salesforce has developed an AI system, “The AI Economist”, that tries to design the optimal tax policy by simulating interactions between multiple individual taxpayers and tax policy makers. In the system, four AI agents work to earn income and trade with each other (i.e. taxpayers), and another AI agent collects tax and redistributes it (i.e. policy maker).
The goal of taxpayers is to maximize individual income by reacting to policy change, while the policy maker to keep optimal balance between productivity and equality by adjusting tax policy. All agents are collectively learning using reinforcement learning without any prior knowledge of tax policy or economic models.
This reminds me of the utilization of COVID-19 modelling and simulation by Australian policy makers to assist and guide the staged policy decision to “flatten the curve”. The models were introduced and explained to the general public in the government news conference in April, which turns out works. This is a real example to showcase the effectiveness of scientific modeling in optimal policy decisions.
Hopefully, in the not too distant future, we could see more and more creative, scientific and data-driven methods adopted by policy makers across the world.
Reference
