Checking the foundations of economics

Some relationships between economic variables are supposed to be stable over time, creating something akin to a law of gravity in economics. But are they really? A new paper tests how solid these foundations of economics really are.

Economic theory assumes that certain “stylized facts” hold true in the long run (whatever “in the long run” means). For example, purchasing power parity assumes that free exchange rates between two currencies should adjust so that a certain amount of money can buy the same goods and services in both countries. That means that the country with higher inflation should see its currency devalue relative to the country with lower inflation. The problem is that in practice, there are transaction costs and frictions. Theoretically, you can fly from Europe to the United States to get a haircut, but the costs are high, so you don’t do it and instead go to your local hairdresser. Hence, exchange rates don’t follow purchasing power parity all the time. Instead, currency valuations have to be quite far away from purchasing power parity to react to this mispricing. And even then, it typically takes years for exchange rates to adjust.

The result of these frictions is that if you test these stylised facts with real life data, you will have a hard time confirming them in practice. If you look at exchange rates over one, five, or even ten years, purchasing power parity seems to be violated all the time. On top of that, there are the real-life problems of what econometrists euphemistically call “failure of cointegration”.

The chart below shows the ratio of private consumption to GDP in the UK going back to 1870. Economists expect that this ratio is constant over time and that consumption rises at the same rate as GDP. In other words, as national income increases, businesses and households have more money and can afford to buy more things. Sometimes, they tend to save a little bit more and consume less, but only to spend those extra savings later (e.g to buy a house or during retirement when income drops). What is saving for other than future consumption? Saving for saving’s sake is a useless exercise.

Consumption to GDP in the UK

Source: Jorda et al. (2017), Macrohistory database

Looking at the chart above shows that over a century or more, the consumption/GDP-ratio seems to fluctuate around one, but there are these nasty episodes when it breaks down altogether. Econometrists may call them “failures of cointegration”, but I prefer the more common term “World Wars”. World wars are regime changes when consumption is curtailed by external factors.

Finally, we have the effects of globalisation where investment and consumption are no longer directed only nationally, but increasingly internationally. Savings can go wherever they find the best investment opportunities and products can be consumed all over the world.

All this makes it very hard to test these stylized facts in real life, but Alexander Chudik from the Federal Reserve Bank in Dallas and his colleagues have developed a new methodology to account for these practical challenges and tested seven of these foundational assumptions in economics for 17 countries going back to 1870:

• Consumption relative to GDP: As the national income grows consumption grows at the same rate since there is no point in saving for saving’s sake.

• Investment relative to GDP: As the national income grows, investments grow at the same rate since lower investments will curtail future growth, bringing the ratio of the two back to 1.

• Imports relative to exports: A country that runs large trade deficits for too long will see its currency depreciate to the point where its exports become so cheap abroad that exports accelerate, and imports decline to reach parity again.

• Debt relative to GDP: Since tax revenues are generated from national income, debt can only be sustained if it can be financed with tax revenues. Lower GDP means that a country can borrow less or has to default.

• Short-term interest rates relative to long-term interest rates: There may be a premium for holding longer-term bonds, but this premium should not change over time, thus creating a stable ratio of short-term to long-term interest rates.

• Inflation relative to long-term interest rates: The difference between long-term interest rates and inflation (i.e. real rates) is constant in the long run and reflects both a risk premium and economic growth.

• Inflation relative to money supply: Increases in the money supply should lead to increases in inflation and vice versa.

These are pretty fundamental assumptions about how an economy works. The first two are basically the reason, why politicians and economists are so obsessed with measuring GDP. It is the main driver of welfare and if you increase GDP growth, you also increase investments and consumption. The last two are fundamental relationships for managing inflation and interest rates. If the growth in money supply is linked to inflation then you want to make sure the money supply grows at a moderate level in order not to create too much or too little inflation, etc.

So, after this lengthy explanation of why testing these stylized facts is so important, yet so difficult, what are the results of this new approach?

The good news is that five out of these seven stylised facts can be confirmed. While looking at them country by country over shorter time spans (even a decade may be too short) does not confirm that they hold, the analysis of the interplay between 17 large economies worldwide over the last 147 years shows that the first five of the stylised facts above all seem to hold true. That’s great because it means our models of how the economy and international trade work are built on sound assumptions.

The worrisome news is that the last two stylised facts (the ones about inflation) do not hold true, even over 147 years. And that is a major problem, particularly for central banks and monetary policy, that I will discuss in more detail tomorrow.