How X is a Treasure Trove for Economists

Aug 10 • Top News • 584 Views • Comments Off on How X is a Treasure Trove for Economists

Elon Musk, the owner of X (formerly known as Twitter), is not a fan of the Federal Reserve. He often criticizes the central bank for raising interest rates. He even tweeted that they could be “the most devastating in history” last December. But the Fed does not mind Musk’s negative comments. In fact, they rather like his platform, as they see it as a reliable barometer of the economy.

X is in a unique position. Its value as a business is questionable, which is why Musk is trying to change that by renaming the company and other strategies. But its value for the economy is a different story. The platform can serve as a useful indicator of both fundamental trends and market sentiment.

X as a Predictor of Market Fluctuations

One way that X can help economists is by predicting short-term changes in stock prices and bond yields. A group of economists, including Francisco Vazquez-Grande, analyzed 4.4 million financial-related posts from 2007 to April 2023. They used a machine learning model to measure the sentiment in each post: A positive one for a soaring stock; a negative one for Musk’s sarcastic remarks about the Fed.

They found that their X financial sentiment index is highly correlated with corporate bond spreads (the difference between corporate and government bond yields that tend to increase as investor pessimism rises). Moreover, posts can not only track financial fluctuations, they can even anticipate them. The sentiment before the stock market opens corresponds to the return on the stock the next day.

In another paper by Clara Vega and colleagues, they found that X sentiment is also closely related to Treasury bond yields. In fact, it is stronger than sentiment indicators from the Fed’s own official communications.

X as a Gauge of Economic Conditions

Another way that X can help economists is by gauging economic conditions. In particular, job loss reports seem to provide timely labor market information. Thomas Keiner and his co-authors created a separate machine learning model to analyze posts with keywords like “job loss” or “layoff notice”. Their job loss indicator reflects official data on employment levels from 2015 to 2023.

The correlation can be very high as government statistics are usually published late and posts appear immediately. For example, X would have detected the decline in employment at the height of the pandemic in 2020 ten days earlier.

X as an Indicator of Monetary Policy

A third way that X can help economists is by indicating monetary policy decisions. Clara Vega and her colleagues believe that X can better predict monetary policy decisions on the day of the announcement than changes in bond yields. Also, the X sentiment index is an effective predictor of shocks in the event of policy tightening, such as rate hikes. Typically, there is frustration in the posts right before these measures.

X does not cause these economic phenomena. It just reflects broader sentiments that are already spreading through the financial markets. But it does provide an additional way to measure such sentiments, which can be quite valuable over time.

Beyond the Fed, some analysts are also finding other potential applications. Agustin Indako of Carnegie Mellon University in Qatar has calculated that post volume alone can account for about three-quarters of the difference in GDP between countries.

Thus, posts, like satellite images of night lights, can help monitor the state of the economy without relying so much on delayed official statistics. This indicator may work best in poorer countries, where strong social media posts reflect the state of telecommunications and smartphone use.

A Trade-off for X

If X is so useful economically, why isn’t it more profitable? The various papers do not explore the gap between X’s struggle for revenue and its apparent usefulness as an economic tool and an information platform. Musk hit on something when he called his platform “an ordinary digital urban area”.

The problem from an economic perspective is that the town square is like a public good such as parks and clean water. Although public goods may be privately owned, it is hard to profit from them, since it is difficult to charge people for all the benefits they provide. Musk is trying to change the economic equation at X by giving extra perks to users who pay $8 a month to verify on the site. Their posts get more promotion and visibility among other benefits. But this requires a trade-off. Paid posts may start to crowd out more meaningful posts from users who do not want to pay. Over time, a platform where money matters more than trust will not function as well as a town square and, as a result, as an economic indicator. A win for X’s finances would be a loss for Fed economists.

Comments are closed.

« »