Roger Bootle here is nominally telling us about how difficult a problem high executive pay is. We don't think that's a problem but that's not our point here. Rather, to examine one of the numbers being used to bolster the argument. We don't think this number is telling us what people think it is:

As Smithers points out, the aggregate data from both the US and the UK tell a worrying tale. In both these countries, fixed investment as a share of GDP has fallen substantially. There are all sorts of possible explanations for this decline but surely a leading candidate is the prevailing structure of executive pay which incentivises business leaders to minimise investment.

Interestingly, this has major implications for the performance of earnings more generally. In America, surely much of the reason for the Trump phenomenon is the extremely disappointing performance of real earnings for the average American. This, too, has several roots. But one is surely US firms’ low investment.

We are, at best, deeply, deeply, unsure about this. We have two reasons floating around in our minds about this.

The first is that what we have traditionally recorded as fixed investment might not be quite what we are recording as it now. We recall one Goldman Sachs report which looked at business investment in software. This was static (as a portion of GDP) over the past decade, decade and a half. So, where is that technological revolution? Fair enough except: investment here is defined as things which are depreciated, written off, over more than one annual accounting period. That's where we get our investment figures in GDP from, corporate accounts. And you may have noted that there's been rather a change in the way that business buys software these days. All that cloud, software as a service stuff. Instead of (at the margin of course) buying proprietary software people are renting structures. To buy a two year licence of Microsoft Office is business investment in software. To sign up to Office 365 is simply monthly running costs. Yet it's (to all intents and purposes) the same software doing the same job.

That is, much of what actually is business investing in software, in computing, is no longer recorded in GDP as business doing so. The size of this difference being usefully indicated by the size of the cloud and Saas businesses out there. That size being large enough to significantly change this line in GDP.

The other is this complaint that extant businesses are paying out ever more of their earnings in dividends and share repurchases, leaving little to reinvest within the corporation. But this is something that we positively desire: we no longer think that the conglomerate is the correct business structure. Further, capital markets are hugely more efficient in allocating capital than they used to be. For example, Nat Rothschild, Bob Diamond and Tony Hayward (with varying levels of success to be sure) have been able to raise large sums from a standing start to go and create new businesses.

That extant, profitable, businesses are paying out those returns back to shareholders, who can then reinvest into other businesses, is not a failure nor an aberration, it's actually something we've long desired that large corporations do.

It is of course possible that CEO pay is a problem, that more investing should be done, as you like. Our point here is that the usual metrics pointed to on this subject don't seem to us to be telling us quite this story. And underlying this is our insistence that yes, economic numbers are indeed important. But the most important part of them is understanding the details of what they're saying. And we're really deeply unconvinced that people are so understanding.