Fellow Economists, Please Explain Sumner’s Analysis to Me
I know it sounds like I’m being sarcastic or trollish, but I mean this sincerely. Over the years I have seen other economists use national income accounting identities to make points about saving and investment, and I am not deft in such things. I think it’s because in my undergrad days I was at an anti-Keynesian / anti-macro school, and then by the time I went to NYU they were teaching formal New Keynesian models with microfoundations and rational expectations, so we didn’t do I+G = S+T kind of stuff.
In that light, then, please read Scott’s recent response to a commenter at EconLog who thought the Fed was distorting the capital markets and causing saving to go into existing assets, rather than new ones. Because the wording might be important, here’s exactly what the guy said: “In a less risk-averse and more economically free environment, savings would be flowing into new investments. But right now all that’s happening is that this savings is being used to bid up prices on existing assets, mainly housing. So this asset bubble is reducing investment in other areas of the economy.”
Now, here’s how Scott apparently blew the guy up:
Saving does not and indeed cannot flow into existing assets, at least in net terms. You might use $X of your income to buy a certain asset, like an existing house. But that’s offset by the fact that the person you buy the house from receives $X for selling the asset. In net terms, no saving has been absorbed in this transaction. In contrast, savings get absorbed when new capital goods are constructed, as when a new home is built. Because saving equals investment, we need never worry about saving being directed into non-investment uses, at least in aggregate terms.
I think I understand (partially) the kind of argument Scott is making, but I don’t think his conclusion follows. Specifically, I think Scott interpreted the commenter to be saying, “Because of the Fed, S > I.” Then Scott was showing that no, that’s impossible, S necessarily equals I.
However, that wasn’t what the guy was actually claiming (go re-read if you need to). In the comments, the guy backed down, no doubt intimidated by Scott’s superior command of economics and his (Scott’s) confidence. But I think the guy should’ve stuck to his guns. Here’s what I would have said:
1) “Scott, you think you’ve proven that savings can flow into a new capital good, but you left out one important thing: When I spend $100,000 on a newly constructed house, that money goes to the seller of the house. So no net saving could’ve been absorbed by this transaction after all–right? I mean, that’s the same argument you used against me.”
2) “Scott, I’m having a hard time reconciling your position with the subjective theory of value. Suppose Smith starts with a $100,000 house. Then he builds another, small house right next to it, and sells that second (new) house for $25,000. We all agree this represents net $25,000 of investment in the economy, right? OK, now suppose instead Smith builds a deck on his original house, and sells the whole thing for $125,000. We all agree that this represents a net $25,000 of new saving and investment in the economy, right? Finally, because all of a sudden Smith is famous, people want to live in his house more than before. His house appreciates in value to $125,000, and Smith sells it to a fanboy. Are you saying this does *not* represent $25,000 in net investment in the economy? Why not? Is it because physical nails and hammers weren’t involved?”
So I am being dead serious, I would appreciate it if someone (preferably a trained economist) could explain to me what Scott meant, and how he would answer the above, hypothetical pushback. (If Scott himself chimes in, even better.)
Yeah this is basically what I was trying to express, but I didn’t feel confident enough to challenge his response. I have a lot of respect for him and what he said seemed not entirely unreasonable, even though it didn’t quite respond to the point I was trying to make.
I’m not a trained economists but when did that ever stop me ?
First of, I agree that Sumner’s statement about money flowing (or not flowing) into existing or new assets is misleading. These transactional flows may affect the pricing of the assets involved but they don’t change the actual stock of money so talk of money flowing into them is wrong whether the transactions are for new or existing assets.
From an accounting perspective some spending in a given period will count as Investment and some as not. The thing that makes it investment is that it is spent on goods produced in that period and not consumed. In all cases money will flow from the buyer to the seller – that has no bearing on the investment v consumption issue.
Some assets just increase in value in a given period with no changes applied to them. Global warming may increase house values in previously cold places. This is clearly an increase in wealth for owners of those asset. I struggle to see how this can count as investment though.
Transformer what is your definition of “investment”? If I am a famous artist and slap some paint on the garage, the market value of my house might go up by $50,000 and I think you would say that is investment, right?
So why is it not investment if the climate changes and the market value of my house goes up by $50,000? Does it really boil down to someone’s arm moving back and forth?
Good point.
I can see the attraction of defining all increases in wealth as being investment.
On the other hand from a flow-of-funds perspective its also nice to distinguish between investment and capital appreciation.
I’m suspecting that as long as you make the definition explicit and use it consistently it doesn’t make any difference to whatever conclusion you may draw.
However I suspect trained accountants have got this one nailed – and it might be smart to stick with whatever they have decided works best.
Based on this I would define investment as “(goods produced in the period and not consumed) – (depreciation of already existing assets) + (appreciation of already existing assets)”.
I think this has the advantage of allowing you to distinguish between gross and net investment.
OK, so now you’ve admitted it’s possible that Sumner’s position is wrong, but you’re not sure where accountants come down on it.
Would you say that accountants think it’s impossible to invest in a previously existing property?
Well, I already said that I think he was misleading with his attempt to distinguish between funds flowing into new investment , while they couldn’t flow into existing investments. In both cases money flows out to sellers just as much as it flows in from the buyers.
But he is still right to reject the commentors point – the fact that fed activity bids up the prices of existing assets has no bearing on the availability of funds for new investment. The bidding up of asset prices destroys neither money nor real resources (at least not directly – it might do do so via the kinds of market distortions Austrians like to talk about in regard to fed policy)
BTW: I think I’m missing how this relates to the capital appreciation point you introduced.
I realize I’m being obstinate Transformer, but it’s because I think you are giving away the game without realizing it.
I’m saying in reality we can’t draw a qualitative distinction between investing in “new capital goods” versus “existing assets,” as different kinds of activities. So it can’t be right to argue that new savings are forbidden from being channeled into existing assets, whereas they can flow into newly constructed ones.
So if you’re with me so far, then you agree there is nothing qualitatively different about a person saving $10,000 and using it to buy a new tractor, versus using it to buy a house for $110,000 that otherwise would have a market value of $100,000.
However, we might think that it’s not good for long-run growth if people are doing the latter, rather than the former.
I disagree with ‘I’m saying in reality we can’t draw a qualitative distinction between investing in “new capital goods” versus “existing assets,” as different kinds of activities’.
If we use the terms strictly then both count as savings but only buying new assets counts as investment since savings can be matched either by dis-savings elsewhere or investment elsewhere.
Here is what I mean.
Assume I earn $10,000. I can spend it on consumption or I can save part or all of it.
If I save by paying someone to build me a new tractor then this is clearly also new investment.
If I save by buying an existing asset then this saving can become either investments or someone else’s dis-saving. That is: The $10,000 I pass onto the asset seller may be used for consumption (my saving is matched by their dis-saving) or by them paying for a new asset to be produced (it becomes investment).
So when you say ‘However, we might think that it’s not good for long-run growth if people are doing the latter, rather than the former.’ I don’t think these are binary alternatives. When I use my savings to buy an existing asset it can still be used further down line line by recipients of the money for investments.
Of course you could argue that when the CB has bid up the prices of existing asset then those selling them at these high values will feel rich and spend more on consumption and less than investment.
If one counts capital appreciation as investment then perhaps in some indirect way then savings by buying old assets (and driving up their price) is creating investment – but I don’t much like that way of looking at it.
BTW Transformer I’m being brief because I’m doing “day job” stuff. You are making good points, I’m not denying that you are doing that. Just saying I think you are basically saying, “Well unless we assume X, this guy had no real point.” But I think the guy was saying X.
Isn’t that the point though? Net savings may not change, but to the extent future investment is redirected because of price changes then we have textbook malinvestment/a bubble.
“Does it really boil down to someone’s arm moving back and forth?”
Well, actually, I think is does. At least, I think that the methods of national income accounting would exclude climate driven capital appreciation from Y if there was no production involved.
Transformer, I’m not sure accounting helps. In an abbreviated form (here with no debt, income, or expense):
assets + appreciation – depreciation
= net equity.
Accounting only cares about the invested dollars and expected return on those dollars.
Perhaps the investment is in the work done by the artist in becoming famous. People do not usually just become famous without doing something, although that seems increasingly to happen. The value of the garage goes up not because the artist slapped some paint on, but because the artist has generated significant value by producing works that are greatly valued.
Transformer, I’m moving our discussion down here to a non-indented part.
You wrote:
“Assume I earn $10,000. I can spend it on consumption or I can save part or all of it.
If I save by paying someone to build me a new tractor then this is clearly also new investment.
If I save by buying an existing asset then this saving can become either investments or someone else’s dis-saving. That is: The $10,000 I pass onto the asset seller may be used for consumption (my saving is matched by their dis-saving) or by them paying for a new asset to be produced (it becomes investment).”
Yes, I think you are re-stating what Scott was trying to do in his EconLog post. But I think you are (like Scott) mixing and matching between two equally permissible lines of argument, which do not yield your desired conclusion when you blend them like this.
Further adding to the irony, I think you anticipate what I’m going to say to you, but you deal with it by saying you “don’t much like” that way of analyzing the situation. Well, I’m sorry if you don’t like logical consistency but… (ha ha)
Anyway, I’m OK with the above (italicized quote from you), but now the obvious problem:
Why can’t the $10,000 I pass on to the person who builds the new tractor be used either for consumption or for saving? It’s the same thing as how you dealt with the seller of the appreciating asset.
Now someone might be tempted to say, “Oh wait Bob, since the guy who builds the tractor is creating new wealth, then that $10,000 payment to him represents net income. So even if he does consume it, he’s not dissaving, he’s just treading water in terms of his own portfolio. So the buyer’s saving is not offset by the tractor seller’s act of consumption.”
Right, but by the same token, if my house appreciates $10,000 during the year, and I sell the whole thing, and then consume only $10,000, then my wealth hasn’t changed either. That capital gain is net income, so there is no dissaving from me to offset the buyer’s saving.
What am I doing wrong in the above analysis? You can “follow the money” around, or you can just consider each financial entity’s balance sheet and income statement, and either way I think you end up agreeing that there is no important economic distinction between buying a new tractor for $10,000 and bidding up the price of an existing asset by $10,000.
EXCEPT, that those different forms of investment might have different effects on the overall structure of production. Just like it makes a difference whether we collectively invest $100,000 in new hammers and $100,000 in new nails, versus investing $200,000 only in new hammers. To explain why the former might be better for long-run GDP growth, we wouldn’t try to argue, “It’s impossible in net terms to only invest in hammers.”
OK, is this what you are saying ?
Define net Investment as any change in wealth in a given period whether its due to new physical assets being produced or external factors like CB actions, climate change etc.
Purchases of existing assets drive up their price , which counts as investment ,- so buying existing assets is just as much investment as buying new assets.
If that is correct then I disagree with the logic.
1. It is not the purchases of the existing assets that drives up the price.- it is the fact that they are seen as more valuable in money terms that causes people to pay more for them. This increase in value is what drive this “investment”: and is external to the purchase decision.
2. From an individual perspective it may make sense to say that the purchase of existing assets is the same as the purchase of new assets – they both have the same effect on the individual’s net wealth. But it does not make sense from the perspective of a closed economy. Investment (as defined above) = increase in wealth = excess of income over consumption + capital appreciation. There will be lots of asset transfers where individuals buy and sell existing assets – but these are is adding nothing to investment and only reflect rather than drive any capital appreciation.
Note: I agree that CB money creation may drive up and distort asset prices and cause people to make bad investment and consumption decisions – I just do not think it is for the reasons I think you are giving here.
Transformer wrote:
“ It is not the purchases of the existing assets that drives up the price.- it is the fact that they are seen as more valuable in money terms that causes people to pay more for them.”
Would you say that purchases of newly produced capital goods drives up their price, or is it the fact that they are seen as valuable in money terms that causes people to spend money on them?
If the latter, then did you just prove people can’t invest in new assets?
The latter. But that just means that for new assets just as much for existing assets capital appreciations is driven by something external to the actual process of buying/selling them , doesn’t it?
There is clear distinction between
– goods that are created but not consumed in a period – these are clearly new investments
– goods produced in prior periods whose ownership is transferred during the period – these are clearly not new investments.(within the context of a closed economy)
But both new and existing assets can have capital appreciation which we are counting as investment.
Transformer wrote:
But both new and existing assets can have capital appreciation which we are counting as investment.
Right, and since Saving = Investment, then you are admitting that theoretically, all of the saving in a given period could be “absorbed” by investment corresponding to the rising prices of existing assets. Right?
And if so, then Sumner is totally wrong. Right?
Yes, since s = i, and investment has been defined to include capital appreciation then it is theoretically possible that all investment and savings may come from this appreciation alone.
But I’m not clear why this makes Sumner totally wrong. He says “we need never worry about saving being directed into non-investment uses, at least in aggregate terms” as even in this extreme scenario this remains true – all savings (in aggregate terms) have indeed become investments.
We might want to worry though that this state of affairs has come about only because people are so happy watching asset prices rise that this becomes their only form of savings – which would be bad news if asset values later fall again.
Transformer, I’m going to have to ask you to go re-read THE TITLE of Scott’s blog post. I’ll wait.
…
Are we good now?
Sorry if I’m being slow or annoying here – but your example shows existing assets realizing capital gains and we defined terms such that this results in both additional savings and investment. to the owners
But I’m not seeing how that scenario shows savings “going into” existing assets – if anything they’re growing out of them.
Transformer, I think we’re moving in a circle here. Every time you bring up a reason to reject what I’m saying, just make sure it wouldn’t equally apply to a new capital project.
So: If a guy puts some parts together and makes a new tractor that has a market value of $25,000, can savings “go into” that newly created asset, or would savings grow out of the creation of the asset instead?
There is nothing magical about moving physical parts around. If you think we can invest in a collection of molecules because some workers rearranged them (i.e. creation of a new capital good), then why can’t we invest in a collection of molecules because we now appraise them differently (i.e. appreciation of pre-existing asset)?
OK, so Sumner calls his post ‘Saving cannot go “into” existing assets’ – and you have shown that if you define investment and savings in such a way that capital appreciation counts as savings – then when you look at where the savings are ‘going’ then some of it is going into existing assets.
I will concede that.
But Sumner may simply reject that way of defining savings to include capital accumulation and then it a semantics debate.
Transformer, I’m not sure what the point of our exchange was, if I was going to lead you step-by-step where you seemed to be agreeing that each move made perfect sense, only to say at the end, “Sumner may simply reject that way of defining saving…and then it is a semantics debate.”
Look, Sumner can write a post saying, “Savings can’t go into new tractors,” and just define spending on tractors as consumption, period. But would that make him right? Couldn’t I write a post saying that is goofy?
I think I get the problem. You are thinking that S = I – C, and so you think spending on pre-existing goods doesn’t correspond to new I, and hence something must be leaking. Right? Something like that? You want to be able to point to the new production to understand how consumption can be less than income?
I think I need to write another post. But I want to make sure I understand exactly what you’re saying. Would you mind, e.g., walking through a simple 2 person economy where 1 guy saves and pays the other guy to build a new house, but then in another scenario the 1 guy saves and buys the other guy’s house for a higher price, and explain to me why one scenario works but the other doesn’t?
I am happy to do that , but before I start may I ask two clarifying questions?
1. Assume we lived in a world where (for whatever reason) assets never appreciated in value and the only source for investment was from not consuming some of the output of the current period. Would you think Scott’s post was wrong in such a world ?
2. My understanding of why you think Scott’s post is wrong is as follows:
When assets appreciate in value you think that this generate both savings and investment. In this scenario savings actually are going into existing assets in contradiction to Scott’s claim that this can never happen.
Is this a correct understanding of your fundamental point ?
1. Yes I agree that if assets can’t go up in price, then it’s impossible for savings to go into existing assets by bidding up their prices.
2. I’m not sure I agree with the claims you are attributing to me. For sure, I think that a capital gain is income, and thus if the owner of an appreciating asset doesn’t raise his consumption and retains the asset, then he has saved and invested in it. But he could sell it to somebody else, in which case that other person invested in it.
my attempt at a model is below (unindented)
The statement in question is completely true:
“In a less risk-averse and more economically free environment, savings would be flowing into new investments. But right now all that’s happening is that this savings is being used to bid up prices on existing assets, mainly housing. So this asset bubble is reducing investment in other areas of the economy.”
A detailed accounting analysis regarding alleged differences between using one’s savings to “invest” in new enterprises as opposed to buying existing assets as a hedge against the inflation that is caused entirely by the funny money emissions that otherwise would not and should not exist is irrelevant to the point being made.
If this were litigation, I’d be seeking sanctions against the opposing attorney for making a frivolous argument. As I’d be doing endlessly against Keynesians and other funny money advocates for trying to cure the problem that does not exist.
I disagree. An accounting analysis using national income accounting rules would show that every time an existing asset changes hands it adds exactly nothing to GDP. This is because GDP accounting rules are designed to measure current period productive output. Moving an item produced in a previous period from one owner to another does nothing to change productive output. Since the flow of money involved in trading existing assets doesn’t show up in Y it also doesn’t show up in C or I. You might argue that Fed policy is giving us too much C and too little I but a) the data for that seems weak https://fred.stlouisfed.org/graph/?g=7XD1 b) that is a totally different argument than “savings are absorbed by existing asset price increases” which at the very least requires totally nonstandard accounting to argue for and c) If we do have all this demand (excessive C) what is it exactly about low interest rates and high equity prices that make it hard for entrepreneurs to raise capital either through debt or equity?
Capt Parker wrote:
Moving an item produced in a previous period from one owner to another does nothing to change productive output.
I get what you’re saying, and I bet even people at the BEA would at first agree with you, but: Did you just prove that businesses that own a fleet of trucks, and ship oranges picked in December 2016 in Florida to New York in January 2017, haven’t done anything productive and don’t count in GDP?
Sorry, Dr. Murphy, I used imprecise wording. Physically moving stuff from a producer to a consumer is productive value adding activity for the purpose of GDP accounting regardless of when the stuff was produced. As you point out real productive capacity gets employed in doing that. Simply transferring ownership after the initial sale of the oranges by the producer is not value added because nothing gets produced due to the transfer and no real productive capacity is utilized simply to make the transfer of ownership. This is true even if the secondhand seller realises a gain. There are lots of fine points: If brokers or middlemen facilitate the transfer and earn a commission then that commission counts as part of GDP. Similarly, retail outfits earn a margin for taking goods from the factory in Texas say, and making them available in a snappy shop in Manhattan. To do this uses real capacity of trucks, drivers, stockgirls storefronts and beancounters so the margin counts toward GDP. GDP accounting excludes non production activity because it is production alone that enables aggregate consumption, whether that consumption be of an orange today or of an orange juice factory that is “consumed” over time.
So what you are saying is that a rising stock market should be ignored, since it’s 99.99% just transfers of existing equities between owners.
This activity is meaningless from an economic sense, right? The stock market serves no purpose then.
I’m saying that if you are talking about I as in Y=C+I+G then the money that goes into buying previously issued equity shares does not count as part of I (or C,Y or G for that matter) If a firm sells new shares which people buy and the firm uses the proceeds of the sale to buy new capital equipment or develop new intellectual property then that expenditure by the firm does count as part of I. It is true that run up in the Dow index say does not automatically mean there is a whole lot of investment happening .
The original comment was mostly correct, although we can’t be certain that prices are being bid up in “mainly housing”. It is true that if someone buys a house, the money is being exchanged from the buyer to the seller. The problem is that if housing prices are being artificially driven higher by government/ central bank policy, then it is a misallocation of resources. It is using up resources (raw materials, labor) to build new houses that should have been put to use in another area that is in accordance with actual consumer demand.
Here is a very simple model that describes the effect of capital appreciation on savings and investment.
Take a closed economy with the following attributes:
– 2 production sectors, investment goods and consumption goods.
– consumption goods have to be consumed in the period they are produced
– investment goods depreciate over many periods and are used to produce consumer goods.
– there are markets for both new and existing goods
-Its a money economy. All trades are in money. All prices are flexible. All markets clear.
– Investment goods and money are the only store of wealth
A the start the economy is in equilibrium. The output of production v consumption goods reflects underlying time preference, and everyone is happy with their level of wealth relative to their savings. Their level of investment exactly matches the rate at which existing investment goods wear out.
There is some trade in existing investment goods between people who want to consume more in the present and those who want to save more, or who want to invest in different sub-sectors of investment goods . The price of investment goods will reflect the current and future stream of consumer goods that they are expected to produce.
Then something happens in the economy that causes the productivity of investment goods to double. (For example: A new production technique is discovered.)
So everyone’s wealth has increased. Price theory shows that the the equilibrium price of investment goods will also increase (in real terms) as now the future stream of consumer goods they will produce has doubled. As all markets clear the new higher price will quickly be arrived at. To an outside observer it may look like new money is pouring into the asset markets and driving up prices during the transition period.
What happens to production?: This will depend upon what the increase in wealth does to people’s underlying preferences. It may be that as people are richer they value leisure and present consumption more and if this is the case the new equilibrium will see less overall production, and a higher ratio of production in consumer goods. But this would be entirely an empirical matter and one could assume what one wants.
What about investment and savings?: In the period when productivity and wealth sharply increases then (based on the definition we have been using) I and S both spike. What happens to investment in new goods is described in the previous paragraph.
Conclusion: When assets prices rise then in the period this occurs some proportion of savings can be attributed to capital appreciation rather than non-consumption out of current production. However in a flexible price model with market clearing there is no sense in which these savings have flowed into “existing assets” to the detriment of “new assets” except when this reflects the underlying preferences of participants in the economy following their increased wealth.
–
It may be that as people are richer they value leisure and present consumption more and if this is the case the new equilibrium will see less overall production,
This statement is a self contradiction. The only way people are made richer from a macro economy standpoint is if the economy is producing more. If overall production is less, Y is less and people are poorer.
Conclusion: When asset prices rise then in the period this occurs some proportion of savings can be attributed to capital appreciation rather than non-consumption out of current production.
If S=I then S=Y-C-G which is an accounting identity. If your intent is to follow national income accounting methods the claim “some portion of savings is from capital appreciation” is wrong. Savings by definition is non-consumption from current production.
What is really going on in your example is:
1) Capital equipment productivity goes up.
2) The price people are willing to pay for capital equipment, either new or secondhand goes up for the reasons you cite
3) Y increases, which is what makes people richer.
4) C. I, and G all may change. How much and is what direction would be an empirical matter and one could assume what one wants provided that:
a) C+I+G= the new larger Y
b) The new I settles in at a value at least big enough to replace worn out capital at the new higher price of the capital and the new production rate.
The increase in the market price of existing capital equipment does not directly change S or I.
Scott is technically correct here for the wrong reason, and I think the confusion is conflating two things. Money isn’t savings, it is the unit of account in these models. Prices reflect the amount of savings/investment, but that doesn’t make the actual money savings and investment.
Consider a lumberjack. He walks into the woods, cuts down some trees and floats them down the river to where his buyer is waiting. Prior to cutting down the trees his savings was 0 (ignoring the ax and human capital) after cutting down the trees his savings is = 10 felled and floated trees. His buyer might pay him 100 lbs of dry beans on the spot, or might give him an IOU for 100 lbs of beans (plus interest for waiting!). Let us look at the net savings in a couple of different scenarios.
1. 100 lbs of beans have been harvested, dried and are stored. In this case the net savings is 10 trees + 100 lbs of beans.
2. The beans don’t exist yet. In this case the savings is = 10 trees + 100 lbs of beans on the IOU – 100 lbs of beans that is now the liability of the buyer, so net savings = 10 trees.
Now let us say that the price of 10 trees increases to 200 lbs of beans, under the two same scenarios
1. 100lbs of beans have been harvested, dried and stored. In this case the net savings = 10 trees + 100 lbs of beans + 200 lbs of beans written on the IOU – 100 lbs of beans that is now a liability for the buyer. Netted out you get the same “savings” of 10 trees + 100 lbs of beans.
2. Same thing here, no beans exists so the net savings = 10 trees + 200 lbs of beans – 200 lbs of beans as a liability = 10 trees worth of savings.
Now we swap in “money” for the IOU. The lumberjack gets payed $100 for 10 trees, and then takes the $100 and buys 100 lbs of beans. Now we know, based on the last market price, that trees are worth $10 each, and beans are worth $1 a lb, with this information we can calculate the dollar value of total savings to be = $200 (even though there is only $100 in existence in this example).
This is great stuff baconbacon and Transformer, reminds me of grad school. I’m about to go on a trip etc. but I hope to get back to this soon. Feel free to bug me in the comments of future posts if I forget.
Thanks Bob.
I certainly have found this discussion very intellectually stimulating.
Transformer and baconbacon, I really do think I get what you are saying. But I’m still not seeing how Scott’s point goes through.
I think Transformer you’re saying something like: If Smith earns a paycheck of $100,00 for his labor, and only consumes $90,000, then he has saved $10,000. Now if he goes to Jones and buys his house, pushing up its price by $10,000 in the process, then we haven’t extinguished $10,000 in savings from the system. Because the very fact that Jones’ house went up $10,000 in price means that *Jones* now enjoys a capital gain of $10,000 and–absent more consumption from Jones–his saving just went up by $10,000. So, Scott Sumner is right, the only way you can “absorb” net saving is to tie it to expenditures on *newly created* capital goods.
But hold on, that can’t be right (Murphy says). Because look at the guy who builds the $10,000 tractor. All of the people involved in selling the materials necessary to assemble the tractor that period collectively earned an income of $10,000. So unless their consumption goes up, this purchase of the tractor corresponds to an increase in saving on their part of $10,000…
See what I mean?
Take a small economy with Fred who makes consumer goods, Smith a wage-laborer , Jones who owns a house, and Tom the tractor builder. All raw materials are free.
Each year Fred pays Smith $100,000 for labor and Tom $10,000 for a tractor. In addition Jones rents a house to Tom for $2000. Fred sells $100,000 of goods to Smith, $8,000 to Tom, $2000 to Jones and consumes whats left himself.
There is clearly $10,000 of saving, right ?
Now one year Smith decides to save $10,000 out of his wages. He buys the house from Jones with the money. Jones sends the $10000 on consumer goods. Fred (that year) sells $90,000 of goods to Smith, $12,000 to Jones, $8000 to Tom. Even though there has been an asset sale nothing changes as far as investment goes – its the same $10,000.
But what if Jones rather than buying consumer goods from Fred, instead buys a newly constructed tractor from Tom for $10,000? In this case we now have $20,000 in savings. If Tom spends the $10,000 he gets on the unused goods from Fred it ends there. But Tom may decide to save some of the revenue in which case there will be additional savings and investment (which will probably be in the form of Fred having unsold goods on his hands). And so on.
So I think only spending on goods that don’t get consumed count as investment. The sale of the house just changes the form of Smith and Jones wealth at the time of the transaction.
What about Jones’ capital appreciation? Suppose he had only valued his house at $5000 at the time he sold it ? He gets $5000 of additional income. If he chooses to buy goods with it that he doesn’t consume – then he’s adding to this economy’s investments.
“So I think only spending on goods that don’t get consumed count as investment.” And capital appreciation of course !
You had it right the first time “only spending on goods (produced in the current period) that don’t get consumed (in the current period) count as investment – period
Jones sold a house for $10,000 so, he has $10,000 to spend on consumption or $10,000 to invest in capital goods or $10,000 to mix and match as he sees fit. This is true if Jones valued the house at $5000, $10,000 $0 or $50,000. And Smith has $10,000 less to consume or invest. So, any increase in investment in the period in which Jones sells his house to Smith has only to do with Jones’ preferences as compared to Smith’s. Jones’ perceived capital gain does not affect macroeconomic investment in any direct manner.
To clarify: I only added that note so I could be consistent with the way Bob is defining investment. But I agree with your conclusions – it makes no difference to the overall story.
Dr. Murphy Said:
So, Scott Sumner is right, the only way you can “absorb” net saving is to tie it to expenditures on *newly created* capital goods.
But hold on, that can’t be right (Murphy says). Because look at the guy who builds the $10,000 tractor. All of the people involved in selling the materials necessary to assemble the tractor that period collectively earned an income of $10,000. So unless their consumption goes up, this purchase of the tractor corresponds to an increase in saving on their part of $10,000…
I’ll hazard answering this a second time in a different way in hopes that I’m more succinct and intelligible this time.
The problem with your objection, Dr. Murphy is that you are looking at the purchase of the newly produced tractor and you are double counting the contribution it makes to the macro economy. Remember that there are two ways to measure the total output of the economy. There is the output-expenditure approach where you measure the total expenditure required to purchase the economy’s final outputs (like oranges or tractors) OR there is the factor income approach where you measure the factor incomes generated by the process of production. Either approach if done right ought to give the same value for national income. You need to use one method or the other not both added together of course.
Now consider the $10,000 tractor. I claim that the sale of that tractor represents a piece of national income, Y, measured by the output-expenditure method to be $10,000 and that since the tractor is not consumed in one period it is an investment (part of I) of $10,000. Since S=I, the $10,000 investment has absorbed $10,000 of saving. End of story. Finished. All she wrote. “But hold on, that can’t be right (Murphy says). Because look at the guy who builds the $10,000 tractor” You can’t do that. If you look at the guy who builds the tractor as you would like to do then, you are trying to count the contribution that the tractor production makes to Y twice – once by the output-expenditure method and once by the factor income approach. The guy who builds the tractor has income, sure, and that income is part of Y and that income can be used in part for C and in part for I but ALL of that income is already accounted for by considering the $10,000 contribution to Y that output-expenditure accounting of the tractor yields.
Sumner is working from a technical definition where “savings” is investment of resources in assets, or something like that, right?
So let’s say we’re in a banana tree economy where someone with more wealth than he wants to invest can:
1) Buy an existing banana tree
2) Plant new banana trees, or
3) Sponsor research into the development of fishing canoes.
The central bank of our village prints up some new money and gives it to me.
a) If I spend the money on an existing banana tree, then so far there hasn’t been any new “savings” an a macro level, at least until we find out where the prior owner spends it.
b) On the other hand, if I spend the money on planting new trees, or if the increased price for trees leads to someone else planting more and consuming less, then there is savings, but arguably the composition of banana savings versus fishing savings is now distorted.
Sumner tends to take people literally – since the commenter said “right now all that’s happening is that this savings is being used to bid up prices on existing assets, mainly housing,” that’s what Sumner addressed. If he had said “regulation and tax policy are directing investment to otherwise unproductive assets like housing,” Scott would have addressed that.
But hold on, that can’t be right (Murphy says). Because look at the guy who builds the $10,000 tractor. All of the people involved in selling the materials necessary to assemble the tractor that period collectively earned an income of $10,000. So unless their consumption goes up, this purchase of the tractor corresponds to an increase in saving on their part of $10,000
GDP, national income C+I+G and all that are means of measuring economic production of final consumption goods and capital goods in the current period. To do this you “follow the money back” until you get to the point where someone is actually producing something. They you stop and mark that transaction down as part of Y. If you don’t stop at the point of production you will be double counting final outputs along with intermediate factors of production.
Example:
Anderson Sells his existing house to Baker for $100,000 – No effect on Y because there was no production keep following the money.
Anderson Buys a car for $30,000 and $70,000 worth of Apple inc. – Mark down $30,000 contribution to Y. Don’t go back any further with that $30,000 because value of all the intermediate factors produced to make the care are already baked in to the price of the car.
Anderson finds a buyer for that Apple stock. and sells it for $100,000 No effect on Y because there was no production keep following the money.
Anderson buys a newly produced house for $100,000 Mark down $100,000 contribution to Y. Don’t go back any further with that $100,000 because value of all the intermediate factors produced to make the house are already baked
into the price of the house.
So the total contribution to Y is $130,000 – $100,000 for the house and $30,000 for the car. Capital gains and secondhand sales don’t contribute because they do not entail production.
In case it’s not clear if a transaction has no effect on Y it has no effect on I ( or S which equals I) since Y=C+I+G
I am not entirely sure your aim in the following so I will take a shot, forgive me if I am off on my interpretation
“But hold on, that can’t be right (Murphy says). Because look at the guy who builds the $10,000 tractor. All of the people involved in selling the materials necessary to assemble the tractor that period collectively earned an income of $10,000. So unless their consumption goes up, this purchase of the tractor corresponds to an increase in saving on their part of $10,000…”
I’m going to change Tractor to Shovel to avoid the complexities of a modern economy. A shovel is a wood handle + a metal blade. Some guy went out and cut a branch off a tree, stripped the bark and shaped it into a handle, then he dug up some ore, melted it and poured it into a mold, and attached the two pieces together, and leans it against the wall of his shop. This is the savings. Some guy comes in and buys the shovel for $10,000 (just for fun!), the $10,000 isn’t “savings”, it is an IOU from someone else, it is net neutral. He is taking an asset of his which is a liability of someone else (in these models the CB or government) and exchanging it for the shovel maker’s savings.
” Some guy went out and cut a branch off a tree, stripped the bark and shaped it into a handle, then he dug up some ore, melted it and poured it into a mold, and attached the two pieces together, and leans it against the wall of his shop. This is the savings.”
Beautifully and simply explained. Thanks something made click when I read your 2 comments!
Thanks!
I think he really is saying that it’s because no nails and hammer were used. If I use wood to build a wheel chair ramp that’s investment in a new ramp. If I burn the wood to roast marshmallows that’s consumption. The first represents savings and investment, and the second does not. S – I is 0 either way.
Sumner is wrong not because he didn’t show an understanding of the alleged effective equivalence between savings bidding up the prices of existing assets and savings being used to buy upgraded existing assets.
He is wrong because he is conflating additional dollars with additional savings (Most monetarists subconsciously do this). For that is the implication of denying that there is an addition to net savings when money is used to invest in an “old” asset, on the basis that the seller of the asset is giving up ownership of that asset, and affirming that there is an addition to net savings when money is spent on a “new” asset (and here he doesn’t seem to understand that in this example there is also a seller who gives up ownership of an asset, which means if the existence of a seller giving up ownership of an asset negates the possibility of net savings taking place, as in the case of the sale of an “old” asset, then no net savings takes place even in the case of the sale of a “new” asset).
It gets even worse though.
Sumner wrote:
“Saving does not and indeed cannot flow into existing assets, at least in net terms. You might use $X of your income to buy a certain asset, like an existing house. But that’s offset by the fact that the person you buy the house from receives $X for selling the asset. In net terms, no saving has been absorbed in this transaction. In contrast, savings get absorbed when new capital goods are constructed, as when a new home is built. Because saving equals investment, we need never worry about saving being directed into non-investment uses, at least in aggregate terms.”
The problem Sumner has created for himself is proposing an inconsequential and irrelevant distinction between “existing” assets and “new” assets. Apparently the fact that an asset was previously sold for money has some bearing on whether the sale of it today for money counts as positive net investment or zero net investment.
If a capital good is purchased today, then we are apparently not able to identify whether a net investment has taken place, until we can find out the history of sales of that same capital good. That is silly. An act is net saving today based on factual and counterfactuals today, not what occurred in the past.
Sumner is also failing to realize that his insistence that savings always equals investment, actually negates his denial that buying an “existing” asset does not constitute an act of net savings. He thinks he is successfully denying it by pointing to the fact that the seller gives up ownership of the asset, but as we have seen above, that is also the case with a seller selling a “new” asset. Sumner’s logic doesn’t work. If the mere existence of sellers giving up ownership of assets negates any possibility of net savings and net investment taking place in those transactions, then Sumner must by his own logic conclude that there is never any net saving or net investment taking place in a monetary economy! Oops.
The existence of continuous positive net investment does not require continuous inflation.
Inflationists typically falsely believe it does for the same reason that motivated Sumner to deny there is net investment taking place in the case of the sale of an “existing” asset. Namely, they only perceive the “offsetting” revenues and costs in dollar flow terms associated with the sale of capital goods. What they don’t take into account is the fact that “costs” for capital assets are depreciated over time. Net investment, if we are going to narrow our focus of it in money terms, very much takes place when an “existing” capital good is sold. The net investment “this period” (net investment always refers to “a” particular period of time) is the difference between the sales revenues and this period’s depreciation costs.
But this cannot be the ultimate reason for net investment. Arbitrary accounting choices must be only part of it. And indeed it is only part. For if we do what Hayek did, and eliminate all monetary changes, that is, if we abstract from money changes, and assume a constant supply of money and constant volume of spending, year are year, then this kind of net investment will eventually fall to zero. It has to mathematically. For at some point, depreciation will rise up to the revenues. The revenues are after all fixed. With continual investment, depreciation will continue to go up until the costs equal the revenues for a given period.
In other words, net investment, in dollar terms (which is a bad way to approach this by the way, it is better to think primarily in real terms) only exists because of inflation. But does this mean growth would stop with zero inflation? On the contrary, growth would be possible indefinitely, on the basis of falling prices.
In real terms, continuous positive net investment takes place on the basis of falling money prices.
Investment is the spending of current period income not on consumption but on things that will increase consumption in future periods. This is the definition that Sumner is using and it is the standard national income accounting definition. If a capital asset that was produced in a prior period changes hands, nothing has happened that will increase consumption in future periods since the productive capacity of that existing asset hasn’t changed simply because of the change in ownership. If current period productive capacity is used to produce a capital asset (that will increase future productive capacity) instead of being used to produce goods and services for consumption in the current period then investment has happened.
Capt. J,
“If a capital asset that was produced in a prior period changes hands, nothing has happened that will increase consumption in future periods since the productive capacity of that existing asset hasn’t changed simply because of the change in ownership.”
Right, but any capital good sold today will always be a capital good “produced in a prior period.” Even what you perceive as the sale of a “new” capital good today, for the first time, is a capital good “produced in a prior period.”
See what is happening here is that you are unintentionally verifying the problem of focusing on dollar flows only. You are putting emphasis on the production of the capital asset. You see a net increase in investment not on the basis of the circumstances of expenditures only, but on the production of capital goods.
We can rewrite what you said above as:
Net investment only takes place when new capital goods come into existence via the production process. Net investment does not take place when an already produced capital good merely changes hands.
You are putting emphasis on the production of the capital asset.
Yes, indeed I am. There is a difference between an economy with only one tractor and an economy with two tractors. To move from an economy with one tractor to an economy with two tractors there must be production of one tractor and hence there must be a net macroeconomic investment equal to one tractor. The economy as a whole must forgo enough consumption to produce the tractor.
To move from an economy with only one tractor that is owned by Adam to an economy with only one tractor owned by Bob there needn’t be production of one tractor and there needn’t be any net macroeconomic investment or any forgone consumption. There isn’t any economic difference between an economy with only one tractor and an economy with (still) only one tractor.
When Bob acquires a tractor from Adam, Bob might think “I have forgone consumption in this period to acquire a tractor so, I have made an investment. But, Bob’s forgone consumption was not forgone by the economy as a whole. Bob’s forgone consumption was transferred to Adam.
Total consumption in a period when the tractor was owned by Bob would be the same as total consumption in a period when the tractor was owned by Adam and this would be the same as total production in the period in which ownership of the tractor was transferred.
Capt J,
In general I think what you are saying is on the right track, however I think some additional caveats or assumptions should be made explicit.
I don’t think you can completely exclude the activity of capital changing hands when making arguments of there being an aggregate or net increase in investment. We cannot just look at the physicality of things. We need to include the value judgments of the buyers.
Suppose that a quantity of capital goods are produced by capital goods producers, but some of those goods go unsold. Suppose they go unsold because they are badly produced. I would argue, and I think you would agree, that we cannot really say there was a net increase in investment just because capital goods were produced. We need to ensure that they are actually valued by individual actors. If the capital goods producers INTENDED to produce those goods not for the purposes of using them themselves, but to sell to others to use in their production process, then the transferring of capital goods from original producers to buyers really must be a component in making the argument that “net investment went up”.
Yes, in a pure physical sense there is no net increase by way of changing ownership, but when you say “investment has increased”, you are tacitly referring to value judgments. That the physical increase in capital goods is associated with an increase in subjective utility or satisfaction. Over the long run, in a free market, yes that tends to be the case, but if we limit our analysis to a given isolated example, it is possible, and in the market it is quite common, for there to be a DECREASE in net investment when there is an increase in capital goods. We would call this “waste”. The standard of reference here is of course a counter factual, but we are in that context already by virtue of comparing capital goods production versus consumer goods production.
Major, You said: I don’t think you can completely exclude the activity of capital changing hands when making arguments of there being an aggregate or net increase in investment. We cannot just look at the physicality of things. We need to include the value judgments of the buyers. … transferring of capital goods from original producers to buyers really must be a component in making the argument that “net investment went up
Yes the price at which ownership is transferred is part of the argument that investment went up. But, neither the transfer itself nor the transfer pricing is part of the creation of the value of the investment. The creation of the value of the investment, whether positive or negative, took place as a consequence of the process of production. The transfer of ownership provided a measurement of the value created by the act of production. Transfer of ownership did not create any investment value by itself.
Here’s an example that I believe supports the idea that only current production can count as investment but addresses the issue of bidding up the price of existing capital: Suppose Apple invents the iPhone 8z. A bunch of geeks say. ” wow, so cool, Apple will make a fortune on those” and they proceed to bid up the price of Apple stock. The investment in this case is Apple’s creation of the iPhone 8z design and the building of an 8z production line. The bidding up of existing Apple stock is not investment. But, the increase in the price of Apple stock on the secondhand market is a measure of the value of the investment Apple made in developing the 8z.
“The creation of the value of the investment, whether positive or negative, took place as a consequence of the process of production. The transfer of ownership provided a measurement of the value created by the act of production. Transfer of ownership did not create any investment value by itself.”
This argument assumes the existence of inherent value, that is, value as a characteristic or predicate of an object, and this value is “measured” by the exchange of it.
This as Mises showed is a problematic:
Murphy wrote an article recently on just this issue:
http://blog.independent.org/2016/08/31/do-individuals-measure-value-ludwig-von-mises-vs-david-friedman/
Major, I’m replying here to avoid infinite indents. Savings and Investment in the context of national income accounting (I.e. in the context in which Sumner was responding pyroseed13) are measures only of objective value as measured by market prices. Sure, both parties benefit from an exchange but, as Mises says the benefit beyond the exchange price is subjective and not measurable. Consequently this immeasurable benefit from exchanges is never included in Y, S, or I. So, you can’t argue that “both sides must benefit from an exchange” means exchanges of existing goods in a secondhand market add to Y, or I and can therefore absorb S. (As I hit submit the Cubs win their first world series in 108 years thanks in part to Theo Epstein who helped break a similar drought for us Red Sox fans)