I am curious about how government purchases crowd out the private sector (if it does). If we look at a graph of the composition of US GDP, it seems that government purchases and investment are somewhat negatively correlated. It's obvious during WWII that government purchases skyrocketed while investment shrank significantly.

I am wondering about the intuition behind why this happens. If I had to guess, it's because at a given snapshot in time, the capital and labor in the economy is limited, and if the government is spending more money to use capital goods and labor, then there are less capital goods to go around for the investment. Is that line of thinking correct?


3 Answers 3


Making the naïve and somewhat incorrect assumption that savings are fixed, every dollar that goes to government spending (e.g. defence, health, education, roads) is a dollar taken away from private investment. This is what is meant by crowding out.

A bit more precisely, we have in a closed economy the following identity: $$Y=C+I+G.$$

We define savings by $$S:=Y-C-G.$$

And hence, $$S=I.$$

So, making the naïve and somewhat incorrect assumption that $S$ is fixed, if $G$ rises by $\$1$, then it must necessarily be that $I$ falls by $\$1$.

The argument above is that sometimes given by opponents of increased government spending.

Supporters of increased government spending counter that the assumption that savings are fixed is naïve and somewhat incorrect. Savings are not fixed, they say. In particular, when $G$ rises by $\$1$, $Y$ may also rise, so that $I$ need not mechanically fall by $\$1$.

But even if this counterargument is correct, there will still probably be at least some crowding out. That is, when $G$ rises by $\$1$, it is likely that $I$ falls by some amount less than $\$1$.

  • $\begingroup$ I think saving is defined as $S:=Y-C-G$, i.e. total national income minus total private and public expenditure. $\endgroup$
    – Herr K.
    Jan 27, 2019 at 4:39
  • $\begingroup$ @HerrK. You are of course right. Now corrected. $\endgroup$
    – user18
    Jan 27, 2019 at 4:46

Just to add some more besides the answers above:

Short answer: by influencing interest rate.

Explanation: In the post Keynesian literature what has come to be appreciated a lot is that government purchases, backed by increased government debt, increases interest rate. Why? Well, that's in itself an interesting explanation and perhaps should be a separate question (interesting, because contrary to what people think that "more govt borrowing means more govt bonds in the market, means lower bond prices and hence higher interest rates" is not at all a satisfactory explanation).

Private investments are assumed to be a function of interest rate. So some of the investment projects (particularly, those at the margin) that were earlier profitable may not be anymore due to increase in interest rate.

This is what is essentially, crowding out of private investment.

This does not require the assumption that Savings have to be fixed. Instead, in Keynesian framework, what happens is that $Y, S, C$ all goes up (because crowding out is not necessarily full, depending on the interest elasticity of money demand - slope of LM curve). It's just the share of $G$ increases and that of $I$ decreases.

Coming to your assessment that historically, at least during some periods, $I$ and $G$ have been found to be positively correlated. Indeed, investment is not just a function of interest rate. It is also a function of sentiments of future of the economy. When govt's start building infrastructure or take measures to improve supply constraints it boosts the confidence in the economy and encourages investment.

That doesn't make concept of crowding out incorrect. Because the concept is applicable when, ceteris paribus, $G$ increases.

Lastly, your own idea that "the capital and labor in the economy is limited, and if the government is spending more money to use capital goods and labor, then there are less capital goods to go around for the investment." is the somewhat closer to the classical view. This assumes that economy is at its full potential and there are no rigidities in the labour market. Without going into details, this view over the year has been seriously challenged. However, it should still be considered as an alternative explanation.

  • $\begingroup$ Federal spending influences the interest rate only as much as the voting members at the central bank want it to. The Fed’s target rate is a policy variable – meaning, it’s whatever they want it to be. dropbox.com/s/759h1yrxyxs2j56/… $\endgroup$ Sep 29, 2020 at 14:36
  • $\begingroup$ One, what happens in US should not be generalized. Two, when federal spending is expanded by financing the deficit, in theory, it means it is accompanied by monetary expansion. So $G$ has not increased ceteris paribus and hence the concept is not directly applicable. Three, if voting members do decide to not monetize the increased spending then also the market interest rates will move. Fed having target rate does not mean interest rate cannot be moved by market forces, but just that Central bank can also influence it. The concept refers to impact of market forces. $\endgroup$
    – Dayne
    Sep 29, 2020 at 15:20
  • $\begingroup$ FOMC members have no part in the decision whether to sell or purchase bonds, as it relates to accompanying new federal spending. They can only decide the target interest rate. The central bank, however, has the infinite capacity to purchase or sell bonds on the open market, so yes, in all but the shortest of terms, they do indeed have the infinite capacity to bend market rates to their will. As I understand it, everything I’ve said, both in these comments and in my above answer, is true for all fully-sovereign currency issuers. $\endgroup$ Sep 29, 2020 at 15:29

Can a fully-financially-sovereign central government use its monopoly power of currency insurance to purchase up a massive amount of real resources, therefore making it such that there are fewer available for the private (non-government) sector to purchase? Yes. That is possible. If you call that crowding out, then crowding out is possible.

If, however you’re talking about any and all government spending “crowding out“ private sector lending based on the assumption of “loanable funds,” then, no, this kind of crowding out does not exist. The idea of loanable funds assumes that there is only a finite pile of money available to both the government and private sector, from which they can “borrow.”

As long as the real resources are available for purchase, federal government spending crowds in (augments) private spending. This idea of crowding out is false and predicated on three other false ideas. Briefly, it incorrectly assumes that:

  1. the central government is a user of its own currency,
  2. banks are revenue/reserve constrained (loanable funds), and
  3. the interest rate is an axiomatic function of the free market.

The truth is that:

  1. the central government issues currency every time it spends (even when it sells bonds, which is very misleadingly called “borrowing”),
  2. banks issue credit every time they lend, and
  3. the interest rate is a policy variable of the Fed’s FOMC board. (They change it because they want to change it.) Further, since the central bank has the infinite capacity to purchase and sell bonds on the open market, they therefore have infinite control over the market interest rate.

In summary: Real crowding out is possible, financial crowding out is not.

  • 2
    $\begingroup$ "the central government issues currency every time it spends," - doesn't it assume that govt always monetizes its deficit - which of course is not true. Most govts borrow to spend beyond its revenues. $\endgroup$
    – Dayne
    Sep 29, 2020 at 13:50
  • $\begingroup$ The revenue from bond sales (very misleadingly called “borrowing”) is not ever collected and then used for future respending. It is destroyed, as all federal revenue is. Bond sales are only used for draining reserves, which is how the central bank manages their target rate. (With a side effect of providing risk-free interest income for the rich, and disemploying millions of poor people.) Bond sales are not at all borrowing in the sense meant by average people. This post contains several academic references: citizensmedia.tv/mmt-securitization $\endgroup$ Sep 29, 2020 at 14:25
  • $\begingroup$ All federal spending is money creation. All federal revenue is money destruction. Bond sales occurring or not occurring don’t change this reality. Bond sales have NEVER changed this reality. $\endgroup$ Sep 29, 2020 at 14:27
  • 1
    $\begingroup$ @Dayne you are actually correct that not all government borrowing creates new money. Aliteralmind is promoting MMT which is fringe and among most economists not accepted theory, that argues that government should fund all its spending with new created money and probably Aliteralmind confused that for saying that all government debt is actually funded that way which it is not in the US. scholar.harvard.edu/files/mankiw/files/… $\endgroup$
    – WilliamT
    Sep 29, 2020 at 16:24
  • 1
    $\begingroup$ @Dayne - I think it's better to engage with the theory to decide on its validity. I think MMT has a lot of useful things to say. I wouldn't just dismiss it out of hand. $\endgroup$
    – Studi
    Mar 22, 2022 at 9:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.