Does Your Model Have HAIR?

We have been following the intellectual struggle between Paul Krugman and the Council of Economic Advisors.  The basic question - can the US stock market continue to produce high returns even in a (projected) period of reduced growth?  The CEA says yes; Krugman says no.

I have spent too much of my time and my reader's patience explaining how the seeming conflict between "return on capital" and growth" can be modeled in the macroeconomic framework employed by the Social Security Trustees.

However, thanks to a very illuminating comment from one of my readers, and drawing on inspiration from Andrew Samwick, I can now resolve the puzzle in a more traditional framework of dividend yield and P/E ratios.

Let's set the stage - folks have been saying, in effect, the following: historically, the US GDP has grown at 3.5% per year, and stocks have had a dividend yield of 3% per year;
profits (net of depreciation) have grown with the economy (and kept a relatively stable share thereof), so, on balance, stocks have produced a real return of 6.5%, composed of 3% dividend yield, and 3.5% capital appreciation, also known as capital gains.

However, folks go on to say, the current yield is (roughly) 1.5%; future growth is forecast to be 2.0%; therefore, we have a problem - if the return on capital is 6.5%, and 1.5% is paid out in dividends, that leaves 5% to be reinvested.  Let's allow 1% for stock buybacks, and we are down to reinvesting 4%.  But if we reinvest (i.e., expand the capital base) at 4%, then profits will grow at 4% while GDP only grows at 2%, and the profit share of US corporations will become an absurdly high proportion of the US economy.  (Here, "absurd" is based on historical evidence, or common sense - over 75 years, a 2% growth differential results in a profit share rising from, let's say, 10% of GDP, up to 44% of GDP.  This is grand news for Bill Gates's heirs, but a bit rough on the rest of us.)

So, folks conclude, since that future world is not plausible, the starting assumptions must be wrong.  But where?  The growth forecast is pretty reliable (based on things like a demographic projection of the workforce), and the dividend yield has an eighty year history behind it.  Aha!  It is the rate of return number that must be suspect.  The world has changed, these people say, and the future rate of return must change with it.  Lower growth, lower returns, night follows day, autumn follows summer, this is the way of the world, and the sun is setting on the US equity market.

Or, somewhat more gloomily, we could have a stock market crash, stocks could start over at greatly reduced prices, and we could march forward with high returns from the new, lower starting point.  But from where stocks are priced now, hopeless!

That, anyway, is the theory as I understand it; if someone would care to guide me to a concise description of such a view, we live to learn.

And we beg to differ.  Let me start with a question - does your model have HAIR?  That would be a Hidden Assumption of Impossible Reinvestment - maybe, in your model, you are simply putting your capital in the wrong place.

Let's switch gears for a moment.  Imagine BigCo has capital equal to 1% of GDP, and provides a 20% return on capital (this must be a huge private equity deal, or folks would bid the price up and the yield down instantly).  My question - can BigCo produce this 20% Return on Capital for the next 30 years?

Some quick noodling tells you that if BigCo grows at 20% while GDP grows at 2% for thirty years, BigCo will grow by a factor of 237 while the economy increases by a factor of 1.8; from its humble origins, BigCo will swamp the US economy.

You, a skeptic, tell me it is impossible.  I salute your facility with exponents.  But you failed - BigCo may well be able to deliver a 30% Return on Capital for thirty years, or forever (absent competitive forces, but don't let's get distracted).

Why?  Your model has HAIR - you made a hidden assumption that BigCo would reinvest all of its Profits (i.e., return on capital) back into the capital base.  As the capital base grew, the profit target grew, and thirty years later, both the profit and capital figures were absurdly high as proportions of GDP.

But why did you assume that?  Suppose BigCo pays out *all* of its profits as dividends.  Then the capital base does not grow, the annual profits do not grow, and BigCo is actually on its way to becoming SmallCo -as GDP grows, BigCo becomes a *declining* share of the US economy.

Hmm.  A 20% return on capital for thirty years with no growth, and it does not swamp the US economy.  Well, if you allow BigCo to pay out dividends, this will work fine.

So, where is the HAIR in our earlier model, with 2% GDP growth, a 2.5% dividend yield, and a 6.5% return on capital?  Well, remember we agreed that the problem with the model was that profits grew to become an unrealistically large number?  We solved that by lowering the rate of return, but we might just as well have raised the dividend yield.

Suppose we had.  Let's picture MyCo with a starting capital of $100, and a return on capital of 6.5%.  First year profits are $6.5; since the economy will only grow at 2%, let's reinvest 2%, or $2, to expand the capital base at that rate.  The rest, $4.5, is paid out as dividends.

Next year, the GDP will be 2% larger; since my capital base is 2% larger, my profits will also be 2% larger, but they will be the same proportion of GDP as previously.  If we continue to pay out a (slightly higher) dividend, and grow the capital base by 2%, we can keep this up forever.  The return on capital will always be 6.5%; the profit as a percent of GDP will not change; and no one will have a problem.

Well, yes, some will have a problem - "isn't that dividend yield awfully high by historical standards"?, someone will ask.  Well, yes.  But I have a quip, and a serious answer.  Quip first?  Dividend policy, like the Constitution, is not a suicide pact.  If companies really don't have any suitable investments available in a low growth economy, it is a bit odd to assume they will invest anyway.  The historic dividend yields reflected a certain historic growth pattern; let the new yields reflect the future!

In a closed system, I think that answer would be compelling.  However, Jim Glass has his hand up, so let me address his point.  This will also reassure those who are desperate to maintain a low dividend yield.

Why did we feel obliged to increase dividends?  Because we wanted to grow the company's capital base (and profits) in line with the growth of the US economy.  But isn't it a big world out there?  Is there some reason that the company can't invest its capital abroad?

Suppose it does.  Imagine that MyCo has $100 of capital, initially all in the US.  It earns $6.5 as before, reinvests $2 in the US as before, but only pays out $1.5 as dividends, for a 1.5% yield.  This leaves $3.  In our original "implausible breakdown" model, this went back into the US, resulting in an implausible level of capital and profits.  But let's invest overseas.

So, year 2:  MyCo has *US* earnings and capital in line with *US* GDP.  No problems - US capital and profits are growing at 2%, in tandem with the economy.

However, a tiny new item appears on the MyCo report - foreign profits of (6.5% x $3) or $0.20.

And from tiny acorns, mighty oaks grow - depending on what one assumes about the reinvestment of foreign and domestic earnings, the foreign side of MyCo will become quite large - in my quick spreadsheet, with dividends paid on both US and foreign capital, reinvestment in the US to keep pace with GDP, and all "excess cash" available after dividends to be reinvested abroad, in 20 years the US capital of MyCo is $148.59, while the foreign capital is $116.74.  (The $148.59 simply reflects 2% growth, obviously).

Now, does this approach lead to a train wreck of implausible proportions?  Not if the world economy grows by 4.5%.  And if it does not, then we fall back to increasing the dividend yield.  Although the total profits of MyCo grow relative to US GDP, the US share reported in GDP accounts does not.  This should be a stable, plausible scenario.

Now, obvious caveats - these numbers and examples do not, and can not, *prove* that the US stock market is correctly valued, or that the dividend yield must remain at 1.5%, or that foreign investment by US corporations *must* grow to some number by some date.

However, what these numbers *do* prove is that a perfectly coherent, mathematically consistent explanation of current share prices, earnings yields, and dividends yields is available.  Toss in foreign investment, and we offer this in refutation of the various "HAIRy" models that attempt to "prove" that stocks must fall.

Folks who would care to vex me are encouraged to bring over some gloom and doom proofs of impending economic calamity, and we will see if we can spot the HAIR on them.

Thanks very much.  I promise my (few remaining) regular readers - we are almost off of this, and we look forward to resuming our regular ranting.

And yes, I still want the Dean Baker "No Economist's Left Behind" Cup.  You haven't seen the end of me!

Comments

[dsquared]

I've posted a lengthy comment on this general area on the previous post which I won't repeat, but just one thing:

therefore, we have a problem - if the return on capital is 6.5%, and 1.5% is paid out in dividends, that leaves 5% to be reinvested. Let's allow 1% for stock buybacks, and we are down to reinvesting 4%.

Noooooo! You have to reinvest the dividends too! If you aren't reinvesting the dividends, then you aren't getting a compound return of 6.5%. You're getting a compound return of 3.5%, plus income of 3%. If that dividend yield isn't compounding, then you can't add it into the total return. So the problem of reinvesting cashflow is just transferred from the company to the investor. The investor has to either buy shares in the market from another US resident (which passes on the problem to somebody else), buy US securities owned by foreigners (which helps a bit, but can't come close to solving the whole problem; not enough of them), or buy newly issued stock (the proceeds of which stock issuance are presumably invested in producing something, which is part of GDP).

The central argument here is the "foreign earnings" one; I agree with you that it's mathematically consistent, although I question its realism in the other comment. This footling around with the dividend payout adds nothing but confusion.

[dsquared]

And you need 6.5% growth in the world economy (more strictly, in the investable world economy), not 4.5%, for the same reason.

(Oh really, lugnuts? How come the USA has had 6.5% returns for a hundred years when it hasn't grown GDP at 6.5%?

Well, there has been an expansion of the PE multiple which it would be foolhardy to extrapolate.)

[spencer]

I can give you the numbers on that too. If you assume that the pe is constant, both earnings and the economy grow at 2% and the dividend yield is 4.5% you soon get to the point that corporations pay out over 100% of earnings in dividends.

All of these scenarios are possible-- a pe of 75, earnings of over 100% of gdp, a divided payout of over 100% of earnings-- in theory. I did not say they were impossible.

It is just more probable that growth would be stronger and in that case there is no SS problem.

The question is, do you want to bet your old age income on these very improbable scenarios?

[Buckaroo]

If Jim G's world, where the US economy shrinks relative to the rest of the world so we're more or less like Sweden today, comes to pass, the US will more or less by definition give up much of it's hegemonic advantage.

This, by itself, would suggest the US would look something like Great Brittain from the 1930s through current day. I believe folks over there were as much concerned about return OF capital as return ON capital during their decline.

Maybe not. Lord knows, how I ever became an advocate for hegemony, I don't know. Maybe there's a world out there where Africa and undeveloped Asia grow at a real rate 4% higher than the west, and we all live happily ever after. Heck, that's sort of what Japan has tried to do - export capital to places like Brazil and the USA and hope it can provide for its aging, slow growth population. Isn't it?

As for raising dividends to a high enough level to compensate - we're talking about compounding gains. You can double dividends from here, but the basic issue is that we're already paying a pretty high price for every dollar of earnings. If we double payouts, there would be precious little capital to reinvest after buying back option shares. You can milk a cash cow for a while and get higher short term returns, but that isn't a formula for compounding growth.

[Jim Glass]

"If you aren't reinvesting the dividends, then you aren't getting a compound return of 6.5%. "

Stocks have never, ever, provided a compound return of 6.5% for any length of time. Why would anyone expect them to do so now?

Stocks have paid an annual return of 7% for 200 years, according to Jeremy Siegel.

Over the 200 years, that annual return has averaged 5% dividends (simple interest) and 2% price growth (compounding). For the last 50 years it's been 3+% dividends and a little less than 4% price growth.

So setting 6.5% compounding as the hurdle to be met seems a tad artificially high.

I hope this isn't the thinking they are following over at DeLong's.

[Patrick R. Sullivan]

"Noooooo! You have to reinvest the dividends too!"

Not if you're retired you don't. Which is the whole point of moving from a 'welfare' system of SS to a true investment system where retirees live off the proceeds from their productive assets.

[Jim Glass]

"If Jim G's world, where the US economy shrinks relative to the rest of the world so we're more or less like Sweden today comes to pass, the US will more or less by definition give up much of it's hegemonic advantage."

As much as goes with moving from 21% of world GDP to about 8%.

BTW, it's your world too, unless you have a plausible scenario for US GDP not dropping as a % of world GDP while US GDP growth drops while the rest of the world's doesn't, because most of the world is young today.

Do you have a scenario for that? Does DeLong?

Compound that 2% annual difference in growth rate that you are all predicting for 75 years, and what do youget?

"This, by itself, would suggest the US would look something like Great Brittain from the 1930s through current day."

Only if we consume our capital financing two world wars, and enjoy a few decades of socialist governments that nationalize major industries and such in the meantime, I'd think.

"Heck, that's sort of what Japan has tried to do - export capital to places like Brazil and the USA and hope it can provide for its aging, slow growth population. Isn't it?"

Japanese corporations invested abroad because they were located in a small island country with few resources and a poor consumer market after the war, and corporate profits were to be earned abroad. Amazingly, it worked!

I just happened to read the history of Honda and there was no mention in it anywhere of its investing in the US market to finance the nation's future retirement liabilities. It was all "big profits = American market, America or Bust!"

Now the Japanese government, which has the duty to finance national retirements, has been running the world's most monstrous deficits of course.

[dsquared]

All the Social Security projections have been based (correctly) on assumptions of compound returns.

By the way, over here in "DeLong World", Japan is an island, but it is not small.

[Jim Glass]

(Oh really, lugnuts? How come the USA has had 6.5% returns for a hundred years when it hasn't grown GDP at 6.5%?

Well, there has been an expansion of the PE multiple which it would be foolhardy to extrapolate.)
~~~~~~~~~~~~

For the stock market to have provided 7% compound returns since 1802 (or even just 1905) in an economy growing 3.5% we'd need a multiple one heck of a lot higher than 20 today! (Just ask Dean Baker.)

So common sense almost tipped you off to your mistake there, dsquared, but not quite.

[TM]

If you aren't reinvesting the dividends, then you aren't getting a compound return of 6.5%.

I will accept Jim's response to that - who said I wanted to, and when did I ever? Look, if this is for my retirement account, I need the money, yes?

Or, as DSquared points out (following Miller-Modigliani), if I want to reinvest, I do. If the revinvestments also return 6.5%, I will personally experience compound growth at 6.5% (try it - set up a model where your company pays a 3% dividend, you reinvest at ever higher prices, and the company's price grows at 3.5%. Tell me your ending account value, and the rate at which it compounded.

If you *don't* get 6.5% (well, subject to tiny rounding), try again - you mismatched the inflows and outflows, or something.

So, multiple responses - it hasn't happened, I may not need it to happen, and with exteranl reinvestment, I can make it happen. (That third is obviously subject to price risk in the real world - so what?).

Otherwise, yes, avoiding the odd results created by compounding is *exactly the point*.

If we double payouts, there would be precious little capital to reinvest after buying back option shares. You can milk a cash cow for a while and get higher short term returns, but that isn't a formula for compounding growth.

Gentlemen - we already agreed that the US economy won't be growing much! Now you are complaining that I am paying out too much in dividends, and not reinvesting enough? I am reinvesting enough to *maintain* the Capital/GDP ratio in a slow growth economy.

Look, the Baker argument is that, subject to restrictions on dividends and a disallowance of foreign investment, the projected cpaital base and profits become laughably high in the US.

Fine, I solved that problem.

Now you say the model is no good because capital is too low? Then maybe the Baker objection is not valid for that reason.

Or if you are saying I should invest more so that we grow faster, even though we can't, well, I'm stuck.

I actually think there is a capital deepening issue, and that the capital ratio should rise over time.

However, that is a mousetrap - in the Dean Baker macro-econ formylation, a rising capital ratio and a steady proft/GDp rati force the return on capital to fall. Yet Krugman said more growth would *raise* the return on capital.

Hmm. I can deepen capital quite easily in my model; Krugman cannot in his, without changing other "assumptions" that give me my result in a low growth world anyway.

[dsquared]

Oddly enough, I nearly added a second point:

"Also, the USA hasn't had 6.5% returns in the sense in which the projections are made. It's had about 4% capital growth which is a compound rate, and the rest is dividend yield. It has been possible for investors during that period to achieve 7% compound by buying more shares with their dividends, but there is no reason so suppose that equity issuance going forward will be so conveniently balanced"

[dsquared]

Or, as DSquared points out (following Miller-Modigliani), if I want to reinvest, I do.

The point here is that this is an option for any one individual, but not (or at least, not obviously) an option for the market as a whole. Up until now, the US stock market has had enough people prepared to sell equities to those who want to reinvest their dividends. If we are planning a big change to our retirement system to base it on the stock market, we should not just assume without checking that buy intentions will match up to sell intentions as neatly in the future. This isn't a small change to the stock market that we're planning here.

You can live off your dividends when you've retired (assuming that you're not required to buy an annuity), but while you're saving, the SS projections are all based on the assumption that you will do something that earns a 6.5% compound rate of return. In order to do that, the money has to be reinvested into productive assets, and these will contribute to the GDP.

[Jim Glass]

"All the Social Security projections have been based (correctly) on assumptions of compound returns."

You mean the Social Security actuaries and Congressional Budget Office independently both make projections "consistent with historical experience" when that experience never occurred?

When one says corporate bonds have had an annual "return" of 4% over the long run, is one saying they have compounded at 4%? If no, then why do you think so with stocks?

Time for a common sense check again:

In 1928 the DJIA was at 240, today it is 10,800.

If it had compounded at 7% more than inflation for those 77 years, how many digits would be in it today and what would its P/E multiple be?

[Jim Glass]

"Oddly enough, I nearly added a second point:

"Also, the USA hasn't had 6.5% returns in the sense in which the projections are made. It's had about 4% capital growth which is a compound rate, and the rest is dividend yield."

Oddly enough, the projections being well based in and consistent with historical experience are based on 4% compounding and 3% dividends -- historical experience!

For anybody to even conceive that 100% of earnings could be reinvested to fuel 6.5% appreciation for some period is what is known as the "Dow 36000 error".

"If we are planning a big change to our retirement system to base it on the stock market, we should not just assume without checking that buy intentions will match up to sell intentions as neatly in the future.

"This isn't a small change to the stock market that we're planning here."

My Gosh, we've already put $11 trillion in IRAs, 401(k)s and such since they were introduced around 1980, far more than is proposed for private SS accounts, and nobody ever thought back then to assure how future stock market buy/sell orders would match up!

The stock market was a lot smaller then too!

Think how we could have screwed it up.

[dsquared]

Jim, the Social Security projections are based on compound rates of return. If you had bought the Dow in 1920 and reinvested dividends, you would have achieved a compound rate of return. The Modigliani-Miller theorem is based on the assumption that you can reinvest dividends to earn a compound rate of return. Nobody else here is confused about compound rates of return. You think you're being clever here but you're not.

[dsquared]

My Gosh, we've already put $11 trillion in IRAs, 401(k)s and such since they were introduced around 1980, far more than is proposed for private SS accounts, and nobody ever thought back then to assure how future stock market buy/sell orders would match up!

Indeed they didn't. And as a result, there was a very considerable expansion of the price/earnings ratio. If your long term assumptions about Social Security include an assumption about future changes in the PE ratio, you ought to make it clear.

[TM]

OK, DSquared's point becoems clear - for Soc Sec account holders, the trustees assumed compound return to estimate future account values.

So, DSq wonders, can that actually be done given the size, the reinvested dividends, etc.?

Still, the funds from Social Security could inject new money into the stock market. The current market value of the Standard & Poor's 500 is about $11 trillion. If 75 percent of those eligible were to opt for market accounts, it could add several more trillion dollars to the market. "It's not a trivial amount," says Mr. Wyss, who estimates it could run as much as 7 percent of the value of the S&P 500.

The new money could come at just the right time, says Taylor. "As boomers retire, there has been a lot of press about the liquidation effect on the market," she says. "But the creation of these voluntary accounts could be sort of offsetting like an intergenerational transfer," she says.

Hmm, a 7% increase beginning in 2009 and phased in over years? I am not worried.

[TM]

Gents, I should not be allowed to have this much fun - I get to bag Krugman and Kos in the same day. Josh Marsahll, some Dem Congressfolks, and the NY Times become collateral damage. They can't pay me enough for this. Stay tuned.

[dsquared]

One point I've perhaps not made clear is that I am perfectly sure that the stock market can deliver these returns; I just think that for consistency's sake, anyone who agrees with me ought to also agree with me that Social Security is not bust under sensible economic assumptions.

You keep on talking about dividends being paid out. But dividends don't disappear when they fall into the hands of savers. They are either consumed or re-invested. In this case, we happen to know that they're not consumed. So they have to be reinvested somewhere, in something; maybe the stockholders use their MSFT dividends to start a hot dog stand or something. And this goes into the GDP.

[TM]

Thank heaven for capital markets and the Miller-Modigliani theorem, which says (among other things) that the dividend decision made by the firm can be re-made by the recipient. Although you question that at an aggregated level.

But if the Soc Sec funds become 7% of the overall market, and those investors tend to reinvest their 5% yeild, that will swamp the market? That is too tiny numbers multiplied together.

[Patrick R. Sullivan]

"Jim, the Social Security projections are based on compound rates of return."

But not for the entire market. There's dis-saving going on too. Especially as the Baby Boomers grow the retirement group. You'll have young workers looking for investment vehicles to purchase, and growing numbers of retirees looking to sell theirs. Everyone won't be reinvesting their returns.

Then there is the rest of the world. Right now Vincente Fox is actively courting investment in Mexico. They're changing their laws governing acquisition of Pacific Coast real estate to allow foreigners to own land in one of the world's most delightful climates. I know people who are looking to develop hotels and golf courses there. It's going to be a brave new investment world.

[Buckaroo]

Patrick, the point is to get the money to compound during the 47 years when you're working.

When you retire, most of the funds will be turned into an annuity. You can hope the company offering the annuity has 6.5% investments with which to give you a reasonable annuity payment. But you won't be collecting many dividends for current consumption on these accounts.

[Buckaroo]

... Japanese government, which has the duty to finance national retirements, has been running the world's most monstrous deficits of course.

Jim, well formed reply. You covered the bases fairly well.

I still don't think we should surrender to a world where the US will hold only 8% of future GDP, rather than the current 21%.

On the way there, when do we stop being the consumer of last resort, and how does that happen?

Would the govt leveraging debt to purchase equity - most of it domestic - really work out well in this sort of world, where our hegemonic position is deterioriating, but the leveraging of debt is stretched out over a 50 year period?

Should we, perhaps, issue BamboozlePalooza Bonds, of 47 year maturity and with interest paid at maturity, so the US could see how things are going and adjust the value of the currency in case we seem to be loosing the bet?

[TM]

My quick answer to DSquared (and I guess, Buckaroo) is that the problem being described is, essentially, the US economy is too small in the furure to support our retirees slice of the GDP pie.

That may be - but the forecast growth is coming form labor force growth and productivity. Saying "gee, raise the reinvetment rate to apply more capital to get more growth" *may* be a sensible solution to the productivity problem.

But my little model is not attmepting to address the problems of the world - I am just refuting Krugman's claim that high returns are "mathematically impossible" with low growth.

Put another way - I may run a great restaurant, that offers a fabulous dining experience for 50 people, and always will. I don't need growth to provide a "fabulous dining experience". But I*do* need growth to offer a "fabulous dining experience" for 75 people.

However, for Krugman to walk over and say, "in the future you can't provide a fabulous dining experience for 75 people; therefore it is mathematically impossible that in the future you will provide a FDEx at all" is wrong.

[TM]

Hmm, now I have a model with HAIR at a restaurant offering a fine dining experience. How is your weekend developing?

[Patrick R. Sullivan]

Well, the first sighted white flag is up at DeLong's blog. The best part being:

"I would like to thank Dean Baker, Barry Eichengreen, Paul Krugman, Tom Maguire, Peter
Orszag, Robert Waldmann, and people who wish to remain off the record for helpful discussions
and comments."

[Jim Glass]

"I still don't think we should surrender to a world where the US will hold only 8% of future GDP, rather than the current 21%."

Well, 8% of GDP would not make the US a third-world country. That's a good deal more than Japan has now.

Also, what really matters is GDP *per capita*. When China passes the US in GDP it will be because it has more than 3x as many people as the US -- but still less than 1/3rd the GDP per capita, which is the measure of welfare.

If the US manages itself wisely it has every opportunity to still be the world's leading, most highly developed, and wealthiest economic power 75 years from now, if only because of the big head start it has on the rest of the world by those measures. Even if its relative size has diminished in gross GDP terms.

[RC]

Tom,
The problem with the scenarios you set out is that they might be plausible for an individual company, but that they become implausible if you generalize them to the economy as a whole.

It's not enough to fiddle with accounting identities to get the numbers you want. Your examples also have to make sense at the macro and the micro level.

In your dividends example, MyCo's share of the GDP stays constant. Now, what if every company acts like MyCo? Then all companies stay at the same relative size, and there's no room for new companies get created. But of course that isn't the way the economy works in real life. Companies go under, new companies get created, and companies grow and shrink in relative size over time. The companies that are growing are likely doing so by investing their profits, or investing by borrowing against future profits. Thus we would expect them to have lower dividend payouts than MyCo.

Moreover, you have to ask whether MyCo's business strategy makes any sense in the context of a competitive market. If MyCo is cutting back on its investment to pay out higher dividends, the likelihood is that one of MyCos' competitors (let's call it OtherCo) will invest a greater share of its profits, and achieve higher levels of productivity that MyCo. OtherCo will thus be able to offer the same product at a lower price than MyCo. In order to match OtherCo's price MyCo will either have to reduce its profit margin, or increase its investment in order to match OtherCo's productivity. Either way, MyCo's dividend payouts go down.

Now let's take the case of international investments. It's true that with increasing globalization, MyCo can gain a share of foreign profits by investing abroad. But it's also true that foreign companies will invest in US and thus gain a share of profits earned in the US. We would expect the profits gained to exceed the profits lost only if the outflow of investment capital from the US to the rest of the world exceeded the inflow. Unfortunately that's not the case. Greater globalization hasn't lead to more net US investment abroad, instead it's lead to a larger share of the world's capital getting sucked into the US. The only way we would expect this relationship to change is if there was a massive increase in US savings, and a concurrent trade surplus. Delong makes basically the same point in more detail in the article he just posted.

So really, Krugman (as is his wont) was overstating his case a bit. Instead of saying that low growth and high returns are "mathematically impossible", he should have said they are "extermely unlikely if you adopt plausible assumptions". But the basic case seems right to me.

[Dick Thomma]

Let me see if I can catch up to you guys:

The easiest way for the stock market to rise at its historical rate is to do what it's done up until now - ride the wave of the growing economy.

Tom MacGuire says another way is for profits to become a larger proportion of the economy.

That can be done by ending all unions, (inviting just enough workers to keep price pressure on wages, but not enough for the economy to grow,) ending taxes on corporations and increasing the tax burden on wages.

Now that we have an idea of how it supposedly can happen, what is the proportion of profit to GDP in the most business-friendly economies in the world? Singapore & similar places?

If the US aimed for and became Singapore by 2050, what would the stock market look like between now and then?

How much does the proportion of capital to gdp have to rise to give the same stock returns we have now but without today's economic growth?

[Dick Thomma]

Instead of Singapore, Krugman likes to say where the Rebublicans want to take us is the United States of 1925 or 1895.

What was the proportion of profits to gdp then?

If that's where the Republicans are taking us, once we get there, we then we expect the stock market to grow in line with the economy - correct? So this is a one-time boost - correct?

How long do we expect this boost to take, and how big do we expect it to be?

[Patrick R. Sullivan]

"Greater globalization hasn't lead to more net US investment abroad, instead it's lead to a larger share of the world's capital getting sucked into the US."

We're talking about a gradual change over several decades, as Baby Boomers retire. Things are going to change.

The Krugman-DeLong-Sawicky Axis...posits the end of capitalism. Talk of 'unlikely'.

Another thing that is being overlooked here is that by changing the incentives facing workers, fewer will retire, or will retire at more advanced ages. Which will increase production (or slow it by less).

[TM]

Thanks for noting that paper, Patrick. I have not looked through the DeLong paper, but getting mentioned was quite a pleasant surprise, and very generous of him.

And from the quick summary, I see thatDSquared came up big with the trade surplus argument. Good job.