A chatty and picturesque discussion of margins

The point of this post is not to pile on after a little dip in the stock market.  I will leave that to CNBC. A week ago, they were highlighting that there was a 100% chance of a 20% rise of equities, because all the strategists were negative. Source: one of the strategists.

My own view is that stocks are roughly fairly priced to deliver quite low returns over the coming years.  The prospect is not exciting enough to be heavily involved, but nor is there a clear case for being short, which is actually novel and irritating.  What I present below is consistent with this take or at least not inconsistent. This post is also very qualitative and lacks precision. You have been warned.

The hypothesis of profit margin “mean reversion” has long been a loser, in my view.  Not only is the “mean” to which margins are meant to “revert” clearly not stable, but advocates of the mean reversion thesis were either unclear of what underlying forces would drive the mean reversion or were oblivious to evidence that those conditions were not in place.

For example, even if margins appear to be “above average”, one ought not expect margins to decline when business confidence has been low, capacity growth is weak, the labor market is loose and we are due for a cyclical bounce of productivity. At some level, the mean reversion theorists needed to just look away from their charts and out the window at what was actually going on.

But a couple years ago, the US cycle entered the stage where a mild compression of margins seemed more likely than a continued widening.  In the event, the margin compression has been a bit steeper than just “mild”, at least when measured using National Accounts data, as I do here.  The cause of this seems to have been dismal productivity performance and steeply declining unemployment, as well as an idiosyncratic shock to the energy patch.

In this note, I relate profits to the cycle and will not focus on technical issues in the energy patch.  However, as a rough pass, it looks like energy accounts for perhaps 50 to 75 bps of the almost 250 bps of margin compression within the domestic nonfinancial corporate sector over the past six quarters.


Some equity strategists believe this effect is now reversing, which would favor a stabilization or even temporary reversal of the recent margin compression evident in the National Accounts data. I have no reason to challenge or confirm that.

Looking forward, there seems little prospect of a durable reversal of this trend, at least within the domestic nonfinancial corporate sector, as measured by the National Accounts data. There is some slippage from those data to the index level performance more familiar to equity strategists.  This reflects, among other things, financials, overseas earnings and accounting differences.  Just briefly on one item, the standard measure of “margins” used by equity strategists has the wrong denominator. They use sales or revenue when should be using value added. But that is a separate discussion.

Economists at Applied Global Macro Research (AGMR) have pointed out that the S&P data tend to hold up a bit better in the late cycle, which is where they presume the US economy now to be.  And I take them at their word on that.

So with those protective caveats in place, what are we looking at here?  As the chart above shows, pretax margins on the domestic operations of US nonfinancial corporations have fallen from 15 ¼% in late 2014 to about 12 ¾% as of the second quarter of this year.  Over this same period, after-tax margins have also fallen about 250 basis points to about 9 ¼%.  This decline is steeper because it is on a lower base (denominator) and reflects that the effective corporate tax rate has stopped declining.

Margins are a far more important determinant of cyclical profit performance than is the corporate top line. Indeed, the latter is important mostly to the extent it drives operating leverage which in turn drives margins. And with margins on a narrowing track. it is no surprise (in an accounting sense) that the level of profits has actually declined 15% over this same period.

Looking forward, it seems likely that margins will erode further, although probably at a slower pace, into the peak of the business cycle.  During the next recession, they will probably decline steeply, although I would guess less steeply than in the last two downturns, and any decline there should be temporary.

If  pretax margins were to narrow 40 bps a year for the next three years, reaching 11 ½% by mid-19, and if the corporate top line were to rise a nominal 4% a year over the same period, then profits growth would be roughly zero.

That crude simulation is almost certain to be off.  In the event, something more dramatic will probably happen. But it strikes me as a reasonable base case that one can tweak as developments arise. And, more to the point, it highlights the importance of the margin cycle.

If margins are going to continue narrowing, even if only slowly, then it will be hard for profits to rise much.  For a finer quantification, I would look elsewhere. As implied by the slug, this post is chatty.  I do this for fun, and for me, neither econometrics nor minutia are fun.


By far, the single most important driver of the margin cycle is the labor share of corporate value added. When it is rising, profit margins are almost invariably falling. It is no coincidence that the labor share has gone up by almost as much as the profit margin has fallen in the past couple years, as indicated by the left panel in chart immediately above.

Please note: for all the charts related to profit margins per se, I set the vertical range at 12 percentage points of corporate value added.  That allows the importance of the variable depicted to be obvious at a glance.

The labor share tends to go up when the labor market is tightening and (more importantly) productivity is underperforming due to cyclical considerations, such as a late cycle surge of hiring relative to output growth.  These conditions are now in place, so the labor share is likely to continue rising, putting downward pressure on margins.

Before, leaving the labor share, I would like to make one point that is not really related to the profit outlook but instead relates to a complaint I often hear from liberal economists and other observers on the left.  They point out the labor share has fallen very steeply since the mid-1990s and that this is “unfair” and contributes to income inequality.

I share the values (and willingness to stick it to the man) of those left-leaning observers, but I am not sure this particular observation is relevant.  The decline in the labor share measured from the last 1990s disappears once you net depreciation out of gross value added and measure the labor share, no-resentment style, as a share of NET corporate value added, which I think more accurately captures underlying economic forces, even as they relate to inequality.

As I read the panel on the right in the chart above, the labor share got unusually high during — and in the wake of — the NASDAQ bubble and “war for talent.” Since then it has fallen very steeply, just as the conventional measure (left panel) shows, but it is not now particularly low.

This is not to say that income inequality is not a real thing. This is most certainly not the Op-Ed page of the WSJ or any other outlet dedicated to denying reality. It is just that the more important drivers of income inequality are elsewhere, and a lot more granular.  Measuring things Marxist style as “labor” and “capital” is probably not very clarifying.


The pressure on margins from a labor share that is rising, even if more slowly than in recent quarters, will probably be reinforced by rising fixed costs, which show up in the National Accounts data as capital consumption (depreciation) and net interest charges.  Logically, depreciation is a function of past investment, the associated rate of growth of the capital stock, and shifts in its composition from between quickly depreciating and slowly depreciating items. I have not seen that modeled very precisely, even by guys who dedicate a lot of efforts, such as AGMR.  But it tends to be cyclical and to continue rising into the subsequent recession, once the cyclical bottom is in place.

Roughly the same point can be made for net interest expense, which is a function of corporate leveraging and the interest rate cycle, both of which are past their sweet spot as regards the influence on margins.  In the chart above, note that I set the vertical scale to have the same range as that for the labor income share.


The final two determinants on the after-tax profit margin, from a national income accounting perspective, are indirect taxes and transfers and then the effective corporate tax rate.  The former are slightly countercyclical from an accounting but also not economically significant. So I will not discuss them, except to say that I airbrushed out the spike of transfers related to the energy patch fines showing up in the Q42015 data. That airbrushing has no effect on the cumulative changes from the peak margin, but it just avoids a confusing blip in the charts.

Finally, the tax rate has stopped going down.  I also don’t have much to offer on that, beyond insisting that a decline from 25% to 0% would be harder to achieve than a decline from 50% to 25%. You don’t need to be a Marxist to see that the zero bound probably applies to corporate taxes too, at least roughly.

One final thought, which is qualitative, just like this entire post.  Historically, profit margin compressions has tended to precede recession, in large part because the rise of the labor share reflects the same cyclical forces that have historically tended to bring a monetary squeeze.

But I don’t think it follows from this that the shift of national income away from “capital” and towards “labor” that is currently ongoing is itself a reason to fear recession. There is no reason, for example, to believe that labor’s consumption propensity is far lower than capital’s.

I would take that on a case by case basis.  And it is probably better to monitor the Fed and any real-economy imbalances you might be concerned about directly, rather than through the prism of the profit share. Personally, I am not particularly alarmed by developments along those fronts, although they are now less unambiguously favorable than they were. In any case, it would seem to be a separate set of discussions.