Finally: The Profit margins debate has been settled

The question of elevated profit margins has occupied investors for quite some time. You know what I am talking about: US corporate profit margins, as measured by national accounts NIPA data, have been above average for the better part of the last decade (see chart below).

CorporateNIPAProfits

The question is whether this margin series is mean reverting or not. The issue is not merely academic, but has serious implications for investors: if mean reversion can be expected, current profits, on average, need to be adjusted downward and stocks would have to be considered expensive. In the opposite case, stocks could be reasonably valued on their current PE rations suggesting a fair value and possibly further upside potential.

Who is right?

Eyeballing the chart suggests there is some mean reversion going on. Even better, mean reversion has some theoretical support: Jeremy Grantham of GMO has consistently pointed out that under a capitalist system this chart has to be mean reverting due to the competition which is attracted by above average margins and returns.

Productivity, the service economy and foreign profits

Opponents of the mean reversion story, such as Bruce Greenwald for instance, argue that “this time it is different” pointing out that there have been productivity gains as a result of the IT revolution and that a larger share of the services sector automatically means higher aggregate margins.

Another line of argument says the high NIPA profit margins are a result of an increasing share of foreign earnings for S&P companies: since profits on foreign sales are reflected in profits (numerator) whereas foreign sales are not counted in GDP (denominator) the figure is automatically skewed upwards, without necessarily suggesting above average profitability.

I have tended to be in the mean reversion camp. After all, it is supported by theory and historical evidence. The increased productivity and higher profits margins arguments never convinced me. Such an increase in productivity should materialize in broadly increased wealth. With lower and lower GDP growth rates and a sputtering recovery and I think this case is hard to make with all the social issues pressing the western world.

The same goes for the allegedly higher structural margins of the services sector: first, I fail to see why the margins of the services sector should be structurally higher. Why shouldn’t there be the same competitive pressures as, say, in the manufacturing sector? Given, that the services sector is less capital-intensive, you could even argue its margins should be lower. Second, it is exactly in the high margin services sector that we see most innovation and competition. We have seen plenty of IPOs of companies that try to compete for Google’s ad revenues. It is not unlikely that sooner or later one or more of them will succeed.

However, as you might have inferred from my use of the word “tended,” I have had doubts.

For one, the “foreign profit” argument has been laid out nicely here. Although, I have a few reservations, it is a well laid out case.

Aggregates are a poor basis for decisions

Further, in good old Austrian tradition, I heavily distrust aggregate figures on which most of the arguments of both camps are based.

As is well-known, national accounts are subject to numerous revisions and distortions – the “foreign profit” camp bases its argument on such a distortion – and to a large part they are based on estimates (imputed rent, deflators…) not market prices. Worse, there have been serious methodological changes over time in how the aggregates are constructed. It doesn’t have to be as extreme as was the case in Nigeria where GDP doubled overnight do to a methodological change, but the effect is there. In my view, this cannot be a reliable basis for any decision. But then, I am not Paul Krugman and do not teach at Princeton 🙂

Unfortunately, aggregate stock index data is not much better suffering from survivorship bias, share count changes and the like. Lots of the companies that generated the historical earnings of the S&P index have been replaced by new companies.

For these reasons, I have always thought that the debate can only be meaningfully settled by a bottom up analysis of the current S&P constituents. Since I was too lazy to do it on my own, I was delighted when I stumbled upon an article titled “Why Jeremy Grantham is Right about Corporate Profit Margins” whose authors did exactly that. This is what they did,

(…)Our data sample consists of 1,079 companies that were members of the S&P 500 index between 1989 and 2013. Given that we are performing our analysis utilizing the components of the S&P 500 index, there is some noise in our calculation introduced as a result of inclusions and exclusions of companies in that index. All data, including fundamentals and price data, are from Factset Global.

We have excluded the financial sector from our analysis, given the significant differences between the income statements of financial businesses versus other businesses, i.e., our analysis uses the S&P 500’s non-financial components. This exclusion means that the yearly average number of companies in our analysis is 408, less than what would otherwise have been the case2. For every calendar year, we utilize the fundamental data of all non-financial companies that were a part of the S&P 500 at the beginning of the year. Data reported anytime in a calendar year is assigned to that calendar year.

All calculations were performed on an aggregate basis. For example, instead of calculating profit margins for every company and then applying a weighting process (e.g., equal weighting or market-cap weighting), we calculate total sales and total profits of the index and derived the index’s profit margin using these totals(…)

I think it makes sense to exclude financials, as their earnings are much less reliable due to the big room management has to massage the numbers especially under abundant liquidity conditions. Further, I would argue that over the long-term financial earnings are a function of the health of the corporate sector anyway. It is difficult to lend profitably if your clients do not earn nice margins themselves. This is what they’ve found,

bottom_UP_ProfitMargin

This settles the “foreign profits” argument once and for all. Yes, profit margins are indeed above average and it has nothing to do with the fact that foreign sales are not counted in GDP.

What about the “higher share of services leads to higher margins” argument?

In order to answer this question the authors have looked at sector margins. Below are the charts for the consumer discretionary and consumer staples sectors.ConsumerDiscretionaryBottomUP

Profit_Cons_Staples

Hmm, both of these non-services sectors sport near record margins. It looks, as if Bruce Greenwald’s view doesn’t hold water. Doesn’t convince you? Let’s look at industrials,

IndustrialsMarginbottomup

Wow, the profit margin of industrials is at a record high.

I consider this topic settled: the higher services share cannot explain aggregate high margins either.

What about higher productivity?

In order to assess whether there have been productivity gains, the authors have the following to say:

(…)As we discussed earlier, one of the reasons offered for the structural shift upwards in profit margins is that technological progress we have made over the last couple of decades and the productivity gains achieved have lowered costs. If this were true, we would see a significant improvement in gross profit margins3 as cost of production per dollar of sales would decline (…)

Sounds logical, here is the chart:

EBIDTA_bottom_UP

Uuups, gross profit margins are not only below average, but have been trending lower for the past decade – no productivity gains to see here.

So what are the reasons for the elevated margins?

The authors have identified the following three factors behind the high margins:

  1. Lower depreciation expense as a percentage of sales
  2. Lower SG&A expense (advertising, R&D, software development…)
  3. Lower interest rate expense

I would probably also add the lack of real wage growth and high government deficits. Government business is usually high margin since you are selling to a dumb, non-economic buyer.

Strikingly, both, depreciation expense as well as SG&A expenses have been falling, which I interpret as a sign that capital consumption in the economy runs high – typical for inflationary periods, such as the past 15 years (I define inflation as an increase in money supply). This is also consistent with record high margins for the consumer staples and discretionary sectors.

Conclusion

This is an excellent analysis that has long been overdue. If no methodological mistakes have been made its insights are extremely valuable and profound.

Although, I have doubts on whether our social order can still be considered capitalist and consider the competitive mechanism seriously hampered, I think profit margins nevertheless will mean revert.* One way, that doesn’t rely on competitive forces, would be via capital consumption which has to show up sooner or later in a reduced standard of living (political unrest) and/or higher interest rates via bankruptcies, once the market realized that the cash flow generating capacity on the asset side has been undermined.

* I sympathize highly with Peter Thiel who argues that innovation occurs mainly in unregulated industries (internet etc.) and argues that the lack of creative change in traditional industries is due to excessive regulation, not because of a lack of potential. Clearly, in a world of increasing regulation and taxes incumbents have a huge advantage over potential challengers.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s