Crowding-Out

After three weeks of writing about tax reform, I decided to give us all a little treat and go back to some data. Below is a graph I did with the help of my Uncle FRED -- the data/graphing service of the St. Louis Fed. That's FRED! https://fred.stlouisfed.org/

What I am attempting to do here is to shed a little light on the impacts of government on investment spending. One worry about tax reform today and government growth generally is something called "crowding-out". Crowding-out refers to the idea that the government and the private sector compete to borrow the nation's saving. If the government borrows a dollar from you, that leaves one dollar less for a private company to borrow. This competition for our saving sometimes crowds out or prevents companies from borrowing, or at least raises the cost of debt enough to curtail spending by companies. 

The chart below plots two data series for the last 56 years. The first, in blue, is annual changes in federal government spending in billions of dollars. The second line is changes in real gross private domestic investment. Real GPDI includes both business investment spending (on plant, equipment, and software) and residential investment. This is the red line.

Before I get into some of the numbers, it helps to put this endeavor in context. This exercise is illustrative and seeks to point out a possible negative connection between government spending and private investment spending. This relationship is admittedly complicated and no single factor ever explains changes in investment. So I am going to prove nothing here. One illustration of my challenge is that the times when investment spending falls the most is during recessions. And those are the times when government spending rises the most. A simple logical error would be to mistake correlation for causation. That is, during recessions government spending rises and investment falls, but this is not the result of crowding-out. It is simply the impacts of a recession on investment and government spending. I won't make that mistake here. 

It is interesting to look into history and see when it appears that weak investment spending was the result of crowding-out by government. This is not going to happen all the time. For one thing, government spending has its own cycles wherein it is sometimes rising, sometimes falling, sometimes rising a lot, sometimes rising a little. We would expect evidence of crowding-out only during in those times when government spending is rising rapidly.

Lots of ifs, ands, and buts. But I still think it is worth the effort. (And what else do I have to do?) Conservatives worry that today's impending tax reform is going to cause government deficits and higher government borrowing, and this will lead to crowding-out. Crowding-out is important. Investment spending is the key to future productivity and economic growth. After the 2008-09 recession, investment spending came roaring back, but you can see in the graph that since around 2012, the changes in real GPDI have decreased and went negative in 2016.

So let's look back so we can think about what might be ahead. The chart starts in 1960 and stretches to 2016. Notice the sharp negative impulses in real GPDI. The biggest decreases were during recessions so I won't discuss those further. What is pertinent for my purpose today are the following episodes:

• After peaking in the mid-1960s, GPDI changes mostly declined through the rest of the decade as government growth was rising.
• After peaking in the early 1970s, GPDI changes declined for several years as government spending changes rose.
• In the late 1990s, government spending growth picked up as GPDI spending got smaller.
• Starting around 2012, government spending increased each year while GPDI spending changes decreased and then turned negative. 
• Finally, I looked hard at this data and can find no durable experiences when rising changes in government expenditures were accompanied by rising changes in GPDI. 

Sometimes government spending increases get smaller. A notable example is the several years starting from the early 1990s. It means the government got out of the way of firms that were trying to raise money in capital markets. That's what we call crowding-in.

• Notice that government spending changes decrease from early 1990s for most of that decade and GPDI changes grew rapidly.

So whether you call it crowding-out or crowding-in, I find six examples that provide some evidence of the relationship between changes in government spending and investment spending. When government spending surges, it tends to limit how much investment can be purchased. When government spending increases decline -- something that does not happen very often -- this improves investment spending. 

Quite clearly, if we eliminate contaminated data surrounding the seven recessions from 1970 to 2016, we are removing much from our study. But focusing on those non-recessionary impacted years, we can see quite a few episodes of government crowding-out and crowding-in. 

Government spending soared after the 2001 recession and during the deeper recession of 2008-09. These increases declined for a while but the end of the chart shows government growing again. And current government spending proposals show no end in sight. The chart shows investment spending tanking as crowding-out would imply. 

Can we prove anything with this kind of analysis? I don't think so. But I think there is plenty of ammo in the data to suggest that if we want to see more national investment in housing, plant, equipment, and software, then we might give some consideration to putting a collar on government spending. My analysis did not examine tax cuts or tax reform that create needs for government borrowing. But I will go out on a limb and argue that any such increases in government borrowing because of lower tax revenues and higher government deficits will not be good for national investment either. Wouldn't it be interesting to be talking about government spending slowing and causing crowding-in!