July 26, 2024
Marketplace radio had a piece on the drop in the saving rate reported for the month of June that committed two major sins. One was ungodly ignorant but inconsequential. The other was more understandable, but far more important.
Starting with the ungodly ignorant one, Marketplace quoted Merrill Reynolds Jr., a faculty member of the banking school at Southern Methodist University, saying people who can afford to are investing rather than saving.
“’It might be real estate, might be stocks,’ he said. ‘Potentially, the stock market or other types of investments out there have certainly been more lucrative — provided you know what you’re doing.’”
This is ungodly ignorant because buying real estate or stocks are both forms of savings. Saving is defined as disposable income (income left over after taxes) minus spending on current consumption. Any money that is not spent on current consumption is by definition saved.
That means that the money that someone might spend on stocks or real estate still counts as savings in the GDP accounts. Back in the old days when I taught intro macro I used to give all the things that counted as saving in the GDP accounts. In addition to buying stocks and real estate, if I bought foreign currency or Bitcoin that would also be forms of savings. If I took my paycheck, cashed it at the bank, and then burned the money, that would still be savings. If I didn’t spend it on current consumption (a newly produced consumption good or service) it is savings.
This matters in this context because if people are using their income to buy stock, real estate, or other financial assets, they are still saving. That cannot explain a drop in the saving rate.
Okay, that is the ungodly ignorant one, now for the more important one. The point of the piece is that the reported saving rate of 3.4 percent is extraordinarily low (see line 35 in the Table). This is true, it was typically in the range of 5.5 to 6.5 percent before the pandemic. But, there is a big problem with the reported saving rate.
This is a somewhat nerdy technical issue, so get your boredom protection shield out. Remember savings is defined as the difference between disposable income and spending on current consumption. There is a huge problem in how these are measured.
In recent quarters there has been a huge gap in measuring GDP on the output side and the income side. This is called the “statistical discrepancy” in the GDP accounts. In the most recent quarter, it was equal to 2.3 percent of GDP (line 34 in the table).
Keep in mind that GDP measured on the output side (consumption, investment, government and net exports) is by definition equal to GDP on the income side (wages and benefits, profits, interest and rents). Think of it like counting bales of hay starting from the left as opposed to starting from the right. There are a certain number of bales in the barn and the number should not be affected by which side you start counting from.
This is the story with GDP. It shouldn’t matter whether we measure it from the output side or the income side, we should come up with the same number. However, in an economy that is nearly $30 trillion, they will never come out exactly equal.
That is ordinarily not a big deal, but the gap has become very large in recent quarters, and it shows output side GDP to be far larger than income side GDP. This matters a lot for the saving rate, which can be easily shown.
First, I don’t know which side is closer to the mark, but it doesn’t matter for this point. Let’s say that the output measure is mostly right, and that income is around 2.3 percent higher than is currently reported.
There are some steps to get from national income to disposable personal income, but any increase in national income is likely to translate pretty closely into an increase in personal income, and since we know taxes pretty well already, it will translate into an increase in disposable income.
Suppose true disposable income is 2.3 percent higher than is now reported. Since consumption has not changed, the true saving rate would be roughly 2.3 percentage points higher than the reported saving rate. That would put it around 5.7 percent, close to the pre-pandemic average.
That’s the story where the output side measure is right, but let’s assume the income side measure is right. In that case, output is 2.3 percentage points lower than is now reported. That error could be anywhere, but consumption is 70 percent of GDP and the government part is pretty well measured, so as a first approximation we can probably assume that if output side GDP is overstated by 2.3 percent, that consumption is overstated by 2.3 percent.
In this case, we get the same result. If the true amount of consumption spending is 2.3 percent below the reported amount, then the saving rate is 2.3 percentage points higher than is currently reported. We are again back to a saving rate of around 5.7 percent, near the pre-pandemic average.
In reality, further revisions to the GDP accounts will likely show the revised number to be somewhere in the middle, but for purposes of calculating the saving rate, it really doesn’t matter which is closer to the mark. Reducing the measured size of the statistical discrepancy will almost certainly lead to a higher measured saving rate that is close to the average in the years just before the pandemic.
In short, Marketplace Radio was trying to explain a problem that does not exist in the world. The drop in the saving rate is an artifact of the data, not something that actually exists.
Comments