Don't forget Tobin's q, Mr Bernanke!

It bugs me when people forget Tobin's q. It's not as though James Tobin was some sort of wild-eyed fringe monetarist. He was a very mainstream Keynesian economist. Ben Bernanke missed an opportunity to invoke the effect of stock prices on Tobin's q and hence on investment as part of his explanation of why loosening monetary policy will work.

Tobin's q is the ratio of the market value of a firm's assets, as measured by its stock and bond prices, divided by the replacement cost of those same assets. It's the value of the bulldozers when they are owned by the firm divided by the cost of buying those same bulldozers. That "average q" is a proxy for the "marginal q", which is what really matters for investment decisions. If buying one extra bulldozer adds more to the financial value of the firm than it costs the firm to buy it, then the firm will increase its stock price by buying the extra bulldozer. And so it should buy the extra bulldozer.

According to James Tobin, one of the main channels of the monetary policy transmission mechanism was through q. Loosening monetary policy would raise stock and bond prices and so raise q and so increase the demand for new investment goods.

Firms can finance new investment by issuing bonds or by issuing stocks. Or by reinvesting earnings on behalf of the stockholders, which is just a shortcut way of issuing new stocks and selling them to the existing stockholders in exchange for those retained earnings, followed by a reverse stock split to keep the number of stocks constant.

If an increased demand for corporate bonds causes bond prices to rise, it makes it cheaper for firms to finance new investment. If an increased demand for stocks causes stock prices to rise, it makes it cheaper for firms to finance new investment. There's an exact parallel between the effect of bond prices and stock prices on investment. If we say that a rise in corporate bond prices caused by monetary loosening will encourage investment, we should also say that a rise in stock prices caused by monetary loosening will encourage investment.

Ben Bernanke talked about the effect of higher bond prices (lower corporate bond rates) encouraging investment. But the only role he gave to higher stock prices was some measly wealth effect and effect on (consumer?) confidence. It almost reads like an afterthought. Stock prices aren't a bit player. They are one of the main actors. What about the effect of higher stock prices on investment? Don't forget Tobin's q!

"This approach eased financial conditions in the past and, so far, looks to be effective again. Stock prices rose and long-term interest rates fell when investors began to anticipate the most recent action. Easier financial conditions will promote economic growth. For example, lower mortgage rates will make housing more affordable and allow more homeowners to refinance. Lower corporate bond rates will encourage investment. And higher stock prices will boost consumer wealth and help increase confidence, which can also spur spending. Increased spending will lead to higher incomes and profits that, in a virtuous circle, will further support economic expansion." (my emphasis).

But I'm so glad to see him invoke the virtuous circle of the multiplier in that last sentence.

Comments

[JKH]

“Or by reinvesting earnings on behalf of the stockholders, which is just a shortcut way of issuing new stocks and selling them to the existing stockholders in exchange for those retained earnings, followed by a reverse stock split to keep the number of stocks constant.”

Retained earnings increase the book value of existing stock. That is unaffected by the market price of the stock, as is the book value of equity in total. The book value is the cash available for investment – not the stock market value.

The comparison you note is that of a firm that pays out 100 per cent of its earnings as dividends instead of retaining the portion in question, with automatic dividend reinvestment instead. That will up the number of shares issued, reduce the book value per share (because there are more shares), but keep the book value of total equity the same - all compared to the straight retained earnings situation. Again, the book value of equity in total is unaffected by that comparison.

So there is no difference in the book value of each shareholders total investment or its market value, comparing those two alternatives for disposition of retained earnings.

There will be an obvious increase in shareholder wealth based on the increase in the market price of the shares.

There will also be an additional economic benefit for the prior existing shareholders if the firm raises new equity externally after the market share price has increased, compared to not doing so after the share price increase. That benefit shows up first as an increase the average book value per share, compared to the alternative of having issued equity before the increase in the market price of the shares. That average book value per share is the actual claim on the investment of cash originally made. And the increase in average book value reflects the initial cash book value benefit that the latecomers (new issue buyers) transfer to the existing shareholders due to the share price increase. That’s how the reduction in the cost of capital shows up as a shareholder benefit when measured in terms of the internal workings of the firm, immediately. Each prior existing shareholder has a greater proportionate claim on the cash (compared to the newcomers) that the firm has deployed in investment, compared to what the situation might have been without both the increase in market price and the new shares issued.

But (I think) the comparative economic effect of Q with regard to new equity issuance is somewhat different than the effect on retained earnings investment. New issuance improves the lot of prior existing shareholders in an additional way that retained earnings does not (relative to the Q effect). Those are the shareholders that are participating in the lower cost of capital. The ones buying the new shares are paying the price.

Since most real investment arguably is financed from new debt issuance and retained earnings, and because the effect on retained earnings is muted, there’s a case to be made that QE has a less direct effect on equity financed investment in total. That’s not to say that the wealth effect on the stock market isn’t important – just that new equity issuance is perhaps not necessarily the major channel. Conversely, the wealth effect due to Q would seem to play only a remote role with regard to the reinvestment of retained earnings.

[Phil Koop]

What JKH said. Plus: while "bit player" may be hyperbolic, in the world of capital markets, equities are small potatoes compared to debt. However, equity plays a disproportionate role in new businesses. So if you are going to talk up the benefits of high equity prices, I think you should focus on IPOs and the related VC financing.

[Nick Rowe]

JKH: I'm not sure I follow. I need to do an arithmetic example so i won't use the wrong words.

Suppose a firm owns 10 bullozers and $100 of cash. It has 10 shares, and no debt. Suppose bulldozers cost $100 new (and don't wear out). So each share represents 1 bulldozer plus $10 cash

Suppose initially the stock market values the future earnings of the bulldozers at $100 each. So each share is priced at $110. Tobin's q is one. If the firm spent the $100 retained earnings to buy one extra bulldozer, and if we assume the stock market would value the earnings of the extra bulldozer at the same $100 as the existing bulldozers (so marginal q is the same as average q) it would have no effect on the share price. So the firm is indifferent to buying the extra bulldozer.

Now suppose stock prices rise, so the stock market values the future earnings of a bulldozer at $200 each. If the firm retains the $100 in cash, each share is priced at $210. But if the firm buys one more bulldozer, each share now represents ownership of 1.1 bulldozers, and the price rises to $220. So the firm will buy the extra bulldozer.

That's what I had in mind.

Phil: sorry. what's VC financing?

[Nick Rowe]

Got it! Venture Capital. Sort of the same as issuing new shares? Only more like an Initial Public Offering, only it's private not public?

[rjs]

who's gonna buy a bulldozer when they've got 6 idle ones parked in the back lot?

[Jon]

What JKH said. Plus: while "bit player" may be hyperbolic, in the world of capital markets, equities are small potatoes compared to debt.

Huh? The notional value of the US Bond market is about 30T, the notional value of the US Stock Market is about 30T. Okay, if you add in M3, you get another 15T (with some double counting), and there are other forms of debt; however, the broadest measures of traded debt amount to only about 70T (distant memory).

Hardly small potatoes: 30T vs 70T.

[Nick Rowe]

rjs: they won't. That's why the *level* of investment is so low. But that's also precisely why the *marginal propensity* to invest is so high. It's why an increase in sales will cause an increase in investment. It's why the marginal propensity to spend will be so high, maybe greater than one. And it's why the multiplier will be so big, maybe even bigger than infinity ;-).

[K]

Phil Koop: "equities are small potatoes compared to debt"

US corporate debt market is about $20Tn.  I'm pretty sure equity market is about $15Tn (Wilshire 5000 is $12.9Tn).  But stocks are way higher beta than corporate bonds (at current spread levels).  Stocks have rallied about 10% since the Sep 21 Fed statement first hinting at more QE.  Corps have rallied maybe 35 bps on average.  Add another 15 bps for the rally in rates.  Assuming a mean 5 yr duration that's a 5*50 bps = 2.5% rally in corps.  (M3 is irrelevant - it doesn't respond at all in terms of market value.) So in terms of reduced asset financing costs: $500Bn from corps, $1.5Tn from stocks.  So about $2Tn of additional Tobin's Q since the past 2 Fed meetings in publicly traded securities alone, 3/4 of it in stocks, and a lot of it arguably the result of QE2.

[JKH]

Nick,

You're example looks straight forward enough.

But under your assumptions, the profit of the firm won't necessarily change from what was the case prior to the stock market increase (or at least not a whole lot except for debt costs - see below). The firm continues to accumulate retained earnings and reinvest in bulldozers. It continues to generate profit - the same level per bulldozer as before - and continues to retain earnings and reinvest etc - and the book value of the firm's equity accumulates with retained earnings, but not much more than it would have been prior to the stock market increase.

The Q is pretty much the same as the multiple of the market value of the stock over book value (pretty much, I think). So that changes the valuation of the firm, relative to book value, and that's all consistent with a higher stock market and a lower cost of capital.

So that's all fine. But given the same profits and same retained earnings and same bulldozer reinvestment, that's not really changing investment demand.

Investment demand will change if additional debt and equity is raised beyond the existing stream of retained earnings.

So I was looking at the difference between retained earnings and new issue stock in that sense, I think.

I.e. how much does the pop in the stock market lead to the kind of new investment that wouldn't have taken place with the existing profit flow? You're example doesn't suggest additional investment that wouldn't have been part of the business plan anyway.

So from there I say that new issue equity is a factor, but maybe new issue debt is a factor more so - because new issue debt would already have been required just to lever up the reinvestment of retained earnings that would have already taken place. The now lower cost of debt does change the assumption about profit levels which also partly accounts for the stock price increase, and maybe that is enough to cause management to expand, etc. But how much? So my point is that the original effect on retained earnings reinvestment is a valuation benefit to equity holders, but is it really a demand effect so much re business expansion?

Rambling response :(

[vjk]

Jon:

the notional value of the US Stock Market is about 30T

What is the source of this number ?

The total stock market capitalization is about $14T.

[The Money Demand Blog (123)]

Bernanke certainly understands the Tobin's q. The problem is that it is a 70 year interest rate, and 5 year interest rates are more important for most investment projects. On the other hand, 70 year interest rates are very important for consumption/saving decisions. If Tobin's q and bond prices are in equilibrium, bond prices are more informative for investment projects.

[Nick Rowe]

JKH: "The Q is pretty much the same as the multiple of the market value of the stock over book value (pretty much, I think)."

Yes. Pretty much. People often use book value as a proxy for the denominator in Q. In my example it would be the same, if the existing bulldozers were bought at the same $100 price. And haven't been depreciated, either in real wear and tear or on the books. There may be some other differences i can't think of. Though it's really *marginal* Q that matters. How much would the stock and bond markets value the earnings of an extra bulldozer as a ratio of the price of buying a new bulldozer.

"But under your assumptions, the profit of the firm won't necessarily change from what was the case prior to the stock market increase (or at least not a whole lot except for debt costs - see below)."

Agreed. (I think QE2 will change those profits, but that's not in my example). I prefer to call those profits "earnings", because they aren't "profits" in the economic sense, and if I call them "profits" I'm going to get an apparent contradiction between assuming the firm maximises "profits" and assuming the firm maximises share price.

The stream of earnings per bulldozer is unaffected by the increase in the price of shares, but the shareholders value that stream of earnings more highly. So the share price increases if the firm takes its retained earnings out of cash and buys bulldozers with it.

Aren't US firms supposed to be "sitting on great wads of cash" right now? And those presumably are retained earnings? If so, my guess is that the main effect of QE2 will be via Tobin's Q on getting firms to spend that cash on new investment.

[K]

TMDB:"If Tobin's q and bond prices are in equilibrium, bond prices are more informative for investment projects."

As discussed above, stocks have rallied 10% over the period of the last two Fed meetings; the 5 yr bond, 15 bps.  Nobody in the real world cares about 15 bps.  Almost entirely due to the stock market rally, US corporations could buy $2Tn of bulldozers and bury them and still not be in a worse situation than than they were in before the start of QE2.  Assuming they have some better ideas than that, I fail to see how a stock market rally is not stimulative.

[Nick Rowe]

TMDB: "The problem is that it is a 70 year interest rate, and 5 year interest rates are more important for most investment projects."

That can't be exactly right. Because firms do issue shares. OK, take a really crude example. Suppose firms insist on having a 60-40 debt equity composition, for God knows what reason, even though bulldozers only last 5 years. If a firm wants to invest $100, they must issue $60 more bonds and $40 more shares. So the cost of financing investment is a 60-40 weighted average of the costs of debt and equity finance.

[Ashwin]

Nick - In your numerical example, why do equity prices go up but not the price of bulldozers? More generally, QE2 has led to an increase in equity prices but also a rise in commodity prices, the same commodities that go into making the capital equipment that firms are supposed to invest in.

If QE results in a fall in real rates and/or increase in inflation, then its logical that all investable real assets increase in nominal value. Right now, almost all non-perishable storable commodities are traded as investments as much as they're traded as raw materials. So the increase in marginal Q may be a lot lower than you think.

[The Money Demand Blog (123)]

Nick, cost of equity has a term structure. Cost of equity for 5 years does not equal to 1/q. It is easier to estimate 5 year cost of equity by adding equity risk premium to cost of debt, rather than subtracting 60 duration risk premium from 1/q.

[Nick Rowe]

Ashwin: the analysis here is about the demand curve for the flow of newly-produced bulldozers. It's the investment-goods demand function. There's also a supply function of newly-produced bulldozers.

In a Keynesian sticky-price macro-model, with excess supply of all goods in a recession, the price of newly-produced bulldozers is sticky, and the quantity produced and sold is determined by the demand side of the market.

In a more Neo-classical model, or when the economy is in equilibrium, a competitive market for newly-produced bulldozers has demand equal supply. If Tobin's Q increases, that shifts the demand curve right, and we get a movement up along the supply curve of bulldozers, and an increase in the price of bulldozers too, if the MC curve of new bulldozers is upwards-sloping.

Some commodities have flexible prices and some also have inelastic supply curves. There we will see a rise in price, but no increase in real investment.

If we make the monetarist/keynesian assumption that the economy as a whole is currently demand-constrained, and suffering excess supply, then we are saying we need to increase demand for goods in general, and the result will be an increase in output. That includes bulldozers.

TMDB: OK. But, under current conditions, does that mean the change in q will over- or under-estimate the 5-year cost of equity?

[K]

Ashwin:"Right now, almost all non-perishable storable commodities are traded as investments as much as they're traded as raw materials."

Bulldozers??  You know that people measure these things, right?  PPI (the Producer Price Index) is stuck below 2%.  Labour costs are the principal determinant of the price of goods.  And there are plenty of good reasons to expect that they aren't going to go up with unemployment at 10%.

TMDB:  Where does 70 (or 60?) years come from.  The duration of a par risk free perpetual bond is less than 25 years.  I can't think of any relevant financial time scale that is much longer than that.

[Nick Rowe]

K: if a stock is like a perpetuity, except that half the coupons are re-invested (say), wouldn't that make it have quite a long duration?

[Jim Rootham]

The stock market is where you go to cash out, not where you go to raise investment cash. That's not absolutely true but it's not a bad first approximation.

The total valuation or the total transactions in the stock and bond market are not good measures of their significance to financing. Look at the total stocks and bonds issued to get a better handle. Also you have to subtract out the money going to founders from IPOs in the stock market to correct for the effects cited in the previous paragraph.

[adjacent / q]

And I thought that the point of high asset prices was to get people to create and sell more assets. No?

[Ashwin]

Nick - Got it. Thanks for the generously detailed reply.

K - Yes, I know PPI is low but we're talking about the determinants of the price of capital goods, not all goods. And I'm not sure labour is the principal determinant of the price of capital goods. Happy to be proved wrong.

Nick - Given my markets background, let me put a different spin on it. Let's assume that the bulldozer doesn't wear out as you just did. In this case, the bulldozer is as much of an investment as the equity is. By a no-arbitrage argument, the price of bulldozers should rise as much as the equity does. I'm not arguing for the sake of it - I get the impression something similar is actually happening in the commodity markets right now.
I can buy sticky prices in consumer goods but sticky prices in capital goods, especially those with significant cost of production coming from booming commodities like metals, I'm not so sure of.

[Determinant]

You can't really say that a stock is a perpetuity because unlike a bond, a stock carries no explicit promise of payment and an explicit promise of risk.

As time approaches infinity, the probability that a firm will go bankrupt approaches one. Therefore the value of a firm is still finite and depends upon its risk profile.

[Nick Rowe]

Ashwin; Yours are good questions (and I can answer them!). It's capital theory you are asking about.

Let's run with your assumptions.

Here's the model that it brings to my mind. There's an existing stock of bulldozers, that are owned directly by people who want a nice income-earning asset from renting them out. They can rent them out for $R per year. Let $P be the price of a bulldozer. (New and used are identical). Assume that R is expected to stay constant (just for simplicity). The Earnings/Price ratio of bulldozers, R/P, is the interest rate on bulldozers. It's their yield. Since the flow of new bulldozers is small relative to the stock of existing bulldozers, there's a perfectly elastic demand curve for new bulldozers at the price P. And the flow of investment in new bulldozers is determined where that perfectly elastic demand curve crosses the upward-sloping supply curve of new bulldozers (the MC curve).

QE2 forces down the required yield on bulldozers, R/P, so P rises. (R is determined by the Value Marginal Product of bulldozers, in the rental market). If P rises, this shifts the demand curve for new bulldozers up. Firms that produce new bulldozers move up along their MC (supply) curves, and investment increases.

Tobin's Q never changes from 1 in that model. But the effect is exactly the same.

That model works quite well for houses, by the way, if you ignore the fact that new houses aren't usually built in the same places as the existing stock of old houses.

It's an old theory of capital. Can't remember who first put it forward, decades ago.

[vjk]

Jim:

The stock market is where you go to cash out, not where you go to raise investment cash.

Right, outstanding shares fluctuations are completely irrelevant for the issuer: his cash won't increase/decrease auto-magically due to QE2 voyage.

To raise cash to buy new bulldozers, he has to issue new shares or bonds.

[Nick Rowe]

Jim, vjk: sure. You have to issue new shares, new bonds, or spend the retained earnings of existing shareholders, to get more investment. I've already dealt with the retained earnings channel. But the prices of new shares and old shares aren't unrelated. Arbitrage should ensure they are the same, for the same shares. And all shares are substitutes for each other. When new shares rise in price, investment increases. the higher are share prices, the more likely it is that you can cover the cost of the new bulldozers by issuing shares.

[RSJ]

K:
"Assuming they have some better ideas than that, I fail to see how a stock market rally is not stimulative."

A long time ago I was interested in this question, and did a long run study of stock movements and re-investment rates during shocks (i.e. the fraction of earnings that were not passed on as dividends). This is because the earnings are controlled by the economy, and the shareholders determine the stock price, so the re-investment rate is the only thing completely controlled by management. The question was, when stocks are re-priced, does the re-investment rate move as well? And my conclusion was ... sometimes. In the 1987 crash, it was as if nothing happened. Management ignored it completely, both on the bull upswing and during the crash. But in the great depression, management liquidated 1-1 with the decline in equity.

Basically management applies a filter and understands that there is a lot of short-term volatility. If you talk to anyone in management they will tell you the same thing: Ask them what their required rate of return is for approving a project. They will not check their stock price, do some division, and give you answer. The required return is a matter of corporate policy, and it moves just as nimbly as you would expect. Target operating margins are also a matter of policy, not CB policy.

Somehow, I don't believe that there are many C-level meetings happening as a result of QE2. But I guarantee you that these meetings happened as a result of the post-Lehman crash.

[Scott Sumner]

This anticipated part of a post I was about to write. Ironically, I asked you about the Tobin q in my reply to your comment over at my blog, not knowing that you had just done a post on the subject. I see this post as a critique of the Krugman post you discussed earlier. Is that right?

[RSJ]

"According to James Tobin, one of the main channels of the monetary policy transmission mechanism was through q. Loosening monetary policy would raise stock and bond prices and so raise q and so increase the demand for new investment goods."

How would you be able to raise prices this way? When the government buys a share, the total number of shares does not decrease, and the dividend payments per share don't change. The earnings prospects don't change. So to the degree that prices reflect fundamentals and not noise, then the price shouldn't change, either.

Again your mechanism only works to the degree that you can force arbitrage to be possible. I.e., a sustained situation in which prices remain above fair value because the CB is able to overwhelm the market. That's not going to happen in complete liquid markets.

[Alan]

Nick,
This may have already been noted in the comments, but QE can also boost q by raising the market value of the real assets (if QE is to have any output effects) that firms invest in. Another mechanism is by lowering risk premia on various assets (here I'm thinking of mortgage loans as a prime example)
Alan

[RSJ]

"Now suppose stock prices rise, so the stock market values the future earnings of a bulldozer at $200 each. "

OK, this is the balloon squeezing assumption, where you are assuming that if the CB buys a lot of shares, then earnings expectations will increase, as opposed to just having the private sector sell their shares to the CB.

Let me try another attack on loanable funds.

There is a pool of those wanting to borrow to invest -- let's call those final borrowers, and a pool of those wanting to defer consumption and purchase financial assets -- let's call those final lenders. These are your demand and supply curves relevant for determining the investment portion of national income. You assume that these two curves meet in a quantity of investment so that both savers and investors are indifferent.

But almost all credit market transactions represent neither final borrowing or final lending, but are merely portfolio shifts. Even though these intermediate transactions cancel out when determining final demand, they do not cancel out in their effect on market prices. It is the relative changes in eagerness of those wanting to buy versus those wanting to sell that moves the market price and causes, say, stocks to decline in value against bonds. Stocks do not decline against bonds because there is a glut of new equity issues and a shortage of new bond issues. They decline in price when the return of bonds is believed to be greater than those of stocks.

The future intentions of the person doing the buying or selling has nothing to do with the effect of the bid on the market price.

Even if we would have 0 final borrowers and 0 final lenders, we would still have market prices of securities, and those prices would be pretty much what they are today. So you cannot argue that the market price is the one where the demand curve of final borrowers meets the supply curve of final lenders. It would be like arguing that the market clearing price of coffee is where the demand of brown eyed coffee buyers is equal to the supply of brown eyed coffee sellers. Even when aggregated, final borrowers and final sellers are just price takers, and cannot move the rate demanded at all even when their own demand curves shift.

Actually, for stocks, net borrowing was negative over the last cycle: firms bought back more shares than they issued -- try to explain that in terms of supply and demand curves involving only final borrowers and final lenders.

Because of this, the market price of the securities may not be the market clearing price if the securities were non-marketable and only final borrowers transacted with final lenders. Then you could argue that asset prices reflect the supply and demand curves of savers and investors, as opposed to portfolio shifters.

And another consequence of this is that you cannot assume that a new entrant will move the price. Suppose the CB announces it will buy stocks at above market-prices. Then you don't just have firms issuing new stocks to the CB, but the portfolio shifters would sell bonds and buy stocks to sell to the CB as well. The price of non-CB purchased assets would fall somewhat and the price of CB-purchased assets would increase.

So the CB would need to purchase all assets at above market prices. And it would need to credibly commit to doing this permanently, because firms are not going to make long range investment plans (and incur long run liabilities) due to a short run financing shift.

But to buy all securities at above market prices permanently means that the government is going to nationalize the productive capacity of the country. If that is on the table, why not just do it and order firms to hire more? Since that isn't on the table, and the CB cannot credibly commit to doing it, then any asset prices swings will be met with front running by the portfolio shifters rather than an increase in investment.

[JKH]

Nick,

“The higher are share prices, the more likely it is that you can cover the cost of the new bulldozers by issuing shares.”

Who’s “you”? It’s the firm, isn’t it? And who’s the firm? It’s the owners, isn’t it? And who are the owners? You have two shareholder groups – one already invested before the rise in Q, and one who’ve bought shares more expensively after the rise in Q.

You say issuing new shares will “cover the cost” of the new bulldozers better than without the rise in share prices. How do you see this applying to these two groups of shareholders? (i.e. to the "you").

[Richard H. Serlin]

It's funny you should bring this up now. I'm in the middle of a long term project to read an intermediate macro book cover to cover, very carefully and thoughtfully. This is partly for review, partly for re-think, and partly because there are sections I never covered as an econ undergrad, or since.

The book I chose was Bernanke's (with Able and Crushore, 6th edition, 2008). And I just recently read his section on investment and Tobin's q. So you'd think he was aware of this at least recently. Quoting from page 138:

Empirically researchers have found that investment in new capital goods does tend to rise when the stock market rises and to fall when the stock market falls, although the relationship isn't always strong...

Although it may seem different, the q theory of investment is very similar to the theory of investment discussed in this chapter. In the theory developed in this chapter we identified three main factors affecting the desired stock of capital: MPKf, the real interest rate, r, and the purchase price of new capital. Each of these factors also affects Tobin's q.

End Quote

Interest rates dropping makes capital cheaper, and a higher stock price means more money can be raised to buy capital from selling stock, but as you said, average q is just a proxy for marginal q, and as Bernanke et. al. said "the relationship isn't always strong". If stocks are up because firms are expecting higher future sales and profits, then it makes sense for them to increase investment. If stocks are up due to a bubble and irrational exuberance, then it makes less sense for them to increase investment.

The question today is how strong is the effect, and will stock prices stay up if it looks like the Fed's move wasn't strong enough and isn't doing much for aggregate demand.

[Richard H. Serlin]

Another thought I had regarding how different "marginal q" can be from average q is that few firms are perfectly competitive with super smooth curves, or very close. Typically there's, to a significant degree, monopolistic competition and steppy curves. Nike's average q will always be far higher than 1, but it's "marginal q" will usually be around 1. Nike's never going to keep investing until their average q goes to 1. It's not profit maximizing given their level of monopoly power. At some point well before average q (or really just q) equals 1, the cost of more capital will be less than the amount it increases stock price, or PV of resulting increased future profits.

In fact, without monopolistic competition or some other deviation from the classical assumptions, I don't think q could be anything besides 1 ever in an efficient market.

[Ashwin]

Nick - once again, thanks for the explanation. It is much appreciated!

So in the context of the supply-demand diagram you just laid out, my point is that as QE results in an increase in price of all investable commodities, the supply/MC curve of bulldozers shifts to the left to the extent that its input costs go up. So for example, if a bulldozer's cost is 70% steel and 30% labour, wages may be unchanged but if the price of steel goes up commensurately the impact on investment via Tobin's Q is correspondingly diminished.

[JBH]

Sir.

"Stock prices aren't a bit player. They are one of the main actors."

Indeed along with housing.

But as you probably already know there are many ways of evaluating stock prices P/E, Tobins Q, Gordon's model( incl.the Fed model ) to name a few.

Unfortunately they seldom yield the same conclusions.

[Panayotis]

A brief comment.

It is not the change in (q) that matters but the change in the present value of expected future earnings relative to (q). Is MP capable to change expected earnings or only the liquidity of finacial markets?

[K]

Ashwin: I think you're right in principle. But in practice I think you vastly overestimate the relevance of commodity inputs. Carbon steel, for example, is $750/tonne. It may only be 10% of the price of a bulldozer.

The PPI surveys prices paid by domestic producers including commodities, stage of production and finished goods (wholesale). Capital goods are part of it. Consumer prices, which are covered by the CPI, are not. But I think there's something else that is being ignored in this discussion. Much of the capital of a firm is human. Replacing it involves hiring and training people. So increased investment will be in the form of labour and not just capital equipment.

Nick and TMDB:

This is how I think about the equity risk premium (could be wrong): The yield on a perpetual is the difference between the discount rate and the coupon growth rate. It is equal to the current coupon divided by the price. For stocks this number is the equity risk premium. The reciprocal of the number is the duration. Since actual payouts vary a lot we don't know the equivalent "current coupon" of the market. But you have to consider both dividends and buy backs. So it's somewhere in between the expected dividend yield (~2%) and E/P (~5%). If we go with 3% we get a duration of 33 years.

[The Money Demand Blog (123)]

Nick said:
"TMDB: OK. But, under current conditions, does that mean the change in q will over- or under-estimate the 5-year cost of equity?"

Good question. I don't have the answer, if I had, I would be writing this comment from a beach in a tropical island. But the 5 year real cost of corporate debt and the 5 year real cost of equity should be in some kind of equilibrium, so under current calm conditions it is likely that change in q will overestimate the change in 5 year cost of equity. Perhaps it was the opposite in March 2009. Anyway, this is the main reason why it is usually a bad idea to do QE with stocks.

K said:
"Where does 70 (or 60?) years come from. The duration of a par risk free perpetual bond is less than 25 years. I can't think of any relevant financial time scale that is much longer than that."

Perhaps it is 50 years, perhaps more, perhaps less. Future dividend policy is a great unknown.

Google has a very low dividend yield, much lower than the perpetual risk free yield.

"But you have to consider both dividends and buy backs."

You also have to consider option grants, not only dividends and buybacks.

[David Pearson]

Nick,

The debt cost of capital is determined real interest rates and credit spreads. Both have been at record lows for some time. The corporate debt markets are fully functional -- corporate debt issuance has been booming. Which sector would you have expected to expand domestically in response to that lower cost of capital, and why didn't they? In global industries, these firms are undergoing in some cases massive expansion -- abroad. In purely domestic industries (for instance housing and commercial real estate), it makes no sense to expand regardless of the Q level. Of course, for the vast bulk if the service sector, the Tobin Q is much less relevant.

The BOJ and MOF engineered several 40% (peak to trough) Nikkei rallies during the two decades-long bear market. None of these sparked sustained investment booms. Its easy to say, in hindsight, "its just because they didn't do enough". At the peak of those 40% rallies, however, I would say market participants were quite optimistic that Japan might enjoy a lasting recovery -- otherwise why did they just bid stock prices up so much? In other words, the "BOJ didn't do enough" because those rallies did not create investment, not because the Tobin's Q never rocketed higher.

[Nick Rowe]

I'm losing it on the comments. Sorry. Too much to get my head around. Just some responses where I think I can make a sensible response:

Richard: I'm using the Canadian edition of the same text. Abel, Bernanke and Kneebone (IIRC). Yes, if you replace r with the desired yield on stocks, the two approaches should be the same.

Here's the way I think about marginal Q: Assume perfect competition, smooth demand and supply curves, and perfectly flexible prices in the market for new bulldozers. In equilibrium, the demand price for new bulldozers always equals the supply price (using those lovely old Marshallian concepts for the height of the demand and supply curves). Marginal Q is the ratio of the demand price to the supply price, so it's always = 1 in equilibrium. But when share prices increase, the demand curve for new bulldozers shifts up vertically, so the demand prices rises relative to the supply price, *at the existing quantity of new bulldozers*. So marginal Q exceeds one *at the old equilibrium quantity*. Quantity then increases, once again closing the gap between demand and supply prices, bringing Q back down to 1. Shifts in curves vs movements along curves. Only instead of thinking of demand and supply curves shifting right and left, and talking about quantity demanded and quantity supplied, in the Walrasian manner, we shift to the Marshallian perspective.

(And the switch from Marshallian to Walrasian perspectives, by the way, explains why economists always draw demand and supply curves the wrong way up, with the independent variable on the vertical axis, which always pisses off my engineering and science students).

[Nick Rowe]

Scott: "I see this post as a critique of the Krugman post you discussed earlier. Is that right?"

I hadn't thought of this post as a response to Paul Krugman. But yes, it would be a partial response. This post was more of a knee-jerk reaction to reading Ben Bernanke's op-ed. "You left out the effect of stock prices on investment, and it's got to be as important as some of that other stuff you are saying to explain why QE2 will work!"

[Chris of Stumptown]

Nick,

It seems like your thought experiment assumes that there is marginal demand for the output of a bulldozer. But if you have a model that assumes that everyone has all the coconuts they need, why should a producer buy more coconut production equipment?

If I am Bernanke, I would say to you that I am not forgetting about Tobin's Q. But the problem is not on the supply side. The problem is on the demand side. I would like to see more done on the fiscal side.

Enough from my sock puppet Ben Bernanke.

I had understood Tobin's Q to be mainly about deciding build versus buy decisions. Is a firm better off purchasing 100 bulldozers, or buying a firm that owns 100 bulldozers?

You are saying capital markets stimulate consumption of capital goods. Probably a fair point. But this neglects any analysis of returns on assets. What happens if more capital goods drives down return on assets in economy as a whole? I am reminded of your "Weird World" post. My answer to that post was that the world now needs bubbles, because ROA is insufficient to provide savers with enough savings.

[Nick Rowe]

Ashwin: remember. Most firms buy a lot of stuff from other firms, so their own value added is only a small part of the cost. But if they increase output in response to an increase in demand, this increases the derived demand for all the inputs they buy. In the final analysis GDP is mostly labour (two-thirds?). Plus, labour can be unemployed or underemployed, but so can all the other resources. And that's as much of a waste and loss of income as unemployed labour. Idle workers, idle machines, same waste.

[Nick Rowe]

Chris: and this goes back to a comment by rjs way back "who would buy a new bulldozer if they have 6 sitting idle?"

If *all* firms had some machines of *all* types idle, then I could see it. But then gross investment would be zero already. And it isn't. So if some firms are already buying new machines, others must be on the margin of doing so. So a rise in shares prices would tip them into buying machines. Which creates demand at other firms, which makes other machines less idle, which increases demand for other machines, and so on. It's not just a multiplier; it's an accelerator. This is what Bernanke meant by the "virtuous circle". Once we add in the accelerator, the existence of slack demand strengthens the effect, rather than weakens it. Investment is part of demand.

I never understood Tobin's Q to be about build or buy decisions. I would see that as a separate question. Whichever is cheaper, build or buy, is what belongs in the denominator of Tobin's Q. And, either build or buy will increase demand for labour and other resources.

[Gizzard]

It seems to me that this may be true in normal times, but right now, in aggregate, it looks as if a rise in stock prices will simply lead to a sell off as folks try to win back some of their losses. Isnt that what happened after QE I? A spike followed by a large selloff? Seems I saw a reference to that somewhere.

You dont make a decision to buy a new tractor cuz its cheap, you make the decision because the tractor will either be sold for a profit OR will allow you to produce something else for a profit. A new tractor at your business will not create people who want to pay you to use it for their benefit. This sounds like "field of dreams" economics, if you build it they will come.

[Chris of Stumptown]

Hi Nick,

Interestingly Alan Greenspan believes the Q argument. I suspect, therefore, that Bernanke does as well. Of course Greenspan's belief in the power of Q got us here so there is the open question of whether higher equity prices are desirable or not. Check out Greenspan from June of 2009:

Global stock markets have rallied so far and so fast this year that it is difficult to imagine they can proceed further at anywhere near their recent pace. But what if, after a correction, they proceeded inexorably higher? That would bolster global balance sheets with large amounts of new equity value and supply banks with the new capital that would allow them to step up lending. Higher share prices would also lead to increased household wealth and spending, and the rising market value of existing corporate assets (proxied by stock prices) relative to their replacement cost would spur new capital investment. Leverage would be materially reduced. A prolonged recovery in global equity prices would thus assist in the lifting of the deflationary forces that still hover over the global economy.

I recognise that I accord a much larger economic role to equity prices than is the conventional wisdom. From my perspective, they are not merely an important leading indicator of global business activity, but a major contributor to that activity, operating primarily through balance sheets. My hypothesis will be tested in the year ahead. If shares fall back to their early spring lows or worse, I would expect the “green shoots” spotted in recent weeks to wither.

source - http://blogs.ft.com/economistsforum/2009/06/inflation-the-real-threat-to-sustained-recovery/

Nick, I think your argument is fine as far as it goes. It reminds me to something I read about Milton Friedman commenting about the helicopter drop of money. In support of its power, Friedman said there would always be a marginal spender to step up as the amount of money dropped grew.

My hesitation is that your model seems so partial. Valuations are only considered against replacement cost, not against returns on assets. What if a capital spending boom drives ROA down in the long term? As Andrew Simthers has repeatedly pointed out, the two valuation measures must eventually converge. It seems like the US may be headed for another Greenspanesque boom which I do not find comforting.

Nick -- is it plausible that prior equity booms have put the US into your "weird world" where savings cannot fund retirements and therefore the economy only works well during a boom? This is one of my working beliefs at present.

[Nick Rowe]

Gizzard: see my comment directly above yours, where I address your point. Some people are already building it, and more will build if it's cheaper.

If I build it, I am coming to the suppliers of the resources I need to build it. And if I come, they will build. That's where the multiplier and accelerator kick in.

Chris: Great find! That's a very clear statement of Q-theory from Alan Greenspan.

"My hesitation is that your model seems so partial. Valuations are only considered against replacement cost, not against returns on assets. What if a capital spending boom drives ROA down in the long term? As Andrew Simthers has repeatedly pointed out, the two valuation measures must eventually converge."

Yes, they should converge, but at a higher rate of investment, and a higher capital stock, than at present. And if people want to save more, that presumably means they want a higher stock of savings, and a higher stock of capital to hold. When they have hit their desired stocks of savings/capital, then they will presumably switch to consumption instead (or decide to stop working, because they have everything they want, which solves the problem of excess supply by reducing supply rather than increasing demand).

[RSJ]

Naked Capitalism has a link to John Hussman's recent market letter about this topic:

http://www.hussmanfunds.com/wmc/wmc101108.htm

"Historically, a 1% increase in the S&P 500 has been associated with a corresponding change in GDP of 0.042% in the same year, 0.035% the next year, and has negative correlations with GDP growth thereafter (sufficient to eliminate any effect on the long-run level of GDP). Now, even if one assumes - counter to reasonable analysis - that the GDP changes are caused by the stock market changes (rather than stocks responding to the economy), the potential benefit to the economy of even a 10% market advance would be to increment GDP growth by less than half of one percent for a two year period.

Now, as of last week, the total capitalization of the U.S. stock market was at about the same as the level as nominal GDP ($14.7 trillion). So a market advance of say, 10% - again, even assuming that stock prices cause GDP - would result in $1.47 trillion of market value, and a cumulative but temporary increment to GDP that works out to $11.3 billion dollars divided over two years. Moreover, even if profits as a share of GDP were to hold at a record high of 8%, and these profits were entirely deliverable to shareholders, the resulting one-time benefit to corporate shareholders would amount to a lump sum of $904 million dollars. In effect, Ben Bernanke is arguing that investors should value a one-time payout of $904 million dollars at $1.47 trillion. Virtuous circle indeed."