Quick point on levelling down

Just a quick point on pensions policy. I've been following the debate on greater compulsion to save for retirement for about 15 years (!). During this time the claim has been repeatedly made by various organisations that complusion, or latterly auto-enrolment, will lead to levelling down.

Typically by this critics mean that while more individuals may save, the average amount of saving may go down (for example, as firms cut back pension contributions to match the minimum required). A further point that has been made is that whilst auto-enrolment may boost pension saving, this could simply represent displacement of other saving. So the overall savings rate might be unaffected.

Well, now we get the chance to test the these predictions (which, needless to say, look a bit like 'Hirschman specials'). My own view, having read a bit about the Aus experience, is that we may well see a bit of shifting of savings (towards pensions, away from others) but we'll also see a net gain. If we don't see an overall increase in saving the critics will have had a very important point proven. But equally if we do then I hope journos who ran 'auto-enrolment will result in levelling down' headlines will go back to those making the claims to get their explanation.

Help me with my maths

Given that in the investment community people only really take an idea seriously if it has some maths to back it up, it must be possible to produce a model showing why proposed 'innovations' in incentive scheme design are likely to deaden motivational effects. As is obvious, I don't think this outcome is a bad one in the short-term, because it hollows out performance-related pay generally. But it would be very valuable to be able to show what the actual effects are.

What I mean by this is as follows. If we are pushing the measurement period for performance awards out, and thus delay the awards being made, then this will mean that recipients discount the value of them (the PwC paper is good on this). If we require executives to hold shares awarded at least for their duration of their period at the company making the award, surely there will be a similar discounting effect. And if we are going to also say that clawback will apply to performance awards too this means that even if awards are made, and sitting there for a number of years before the recipient can cash them in, they are not actually a "sure thing". Again, this must make them discount the value of the award? In combination this could be pretty significant (again, PwC put some numbers on discounting, and they are pretty big).

It must be possible for someone to set out a relatively simple equation explaining all this? I am amazed that people seem to be unthinkingly accepting the proposition that long-term awards combined with longer holding periods 'solve' the incentive problem and create 'alignment'. As spelled out in a previous post, I think what the combination will actually do is nullify behavioural/motivational effects (which is fine, but then why make the awards in the first place). But it would be very useful to set this out in a more formal way.

Any ideas?

Andy Haldane on corporate corporate governance

Will Davies has a great interview with Andy Haldane up on Open Democracy. Go and read it. Here's a chunk on corp gov.

AH: Any institution – it could be a bank, it could be a building society, it could be a non-financial corporate, it could be the Bank of England, any decision-making body – is likely to benefit from a greater plurality of thinking around the table. It could be from different experience or from a different intellectual toolkit. I think the experimental evidence on that is very strong. But this plurality does come at a cost. And the cost is that it may on average lead to slightly slower, slightly ‘inefficient’ decision-making. Because it then has to be slightly more consensual and the scope for decisive action is slightly diminished. But that is more than offset by the benefits of not making catastrophic errors. What you lose on the swings you more than make up for on the roundabouts, when you avoid catastrophic errors, which are much more likely in non-plural governance settings. So I think the case in favour of plurality is very strong. 

My second point, which is slightly more specific to finance, is that we have seen a particular form of singular, non-plural decision-making emerge. In UK Company law, the primary responsibility of management is to shareholders. And we’ve also seen managers of those firms being remunerated in a form which aligns their interests with those of the shareholders through payment in equity or equity-like instruments.

Within banking and finance, this has led to a corporate governance structure in which those owning maybe 5% of the balance sheet – i.e. the shareholders – have the primary, some would say the exclusive, power in controlling the fortunes of the firm. There is no say from the debt-holders or depositors or workers or any sense of the wider public good which we know to be important in banking and finance. We also know that those firms are working on time horizons which in some cases are really quite short. So to think that this will necessarily lead to the best outcome, even for the longer-term value of the firm, is questionable given the governance model. 

The piece I did last year was to try and understand how it was that banks ended up with the corporate governance structure I’ve described. What was it? At each stage it was a sensible reason. But it led to a corporate governance structure that looks pretty peculiar, given where we started off 150 years ago. So what I mentioned about structures and incentives – an important thing about that is who runs the firm and how they run the firm. I think corporate governance in the way I’ve defined it is super important - more important than regulation in getting us into a better place.

Corporate charges for phone hacking?

An interesting piece appeared on The Guardian website tonight about possible charges against former News International directors under the Regulation of Investigatory Powers Act (RIPA). Section 79 of RIPA (below) specifically refers to the criminal liability of directors relating to crimes committed by the "body corporate". Section 79 got mentioned in a few news stories when the hacking scandal exploded last July, but there hasn't been a lot of focus on it since. Then the culpability of News International directors did briefly get an oblique reference in Sue Akers evidence right at the end of the Leveson Inquiry hearings.

I suppose the three people you might think most at risk, in descending order are Rebekah Brooks, Les Hinton and James Murdoch. Brooks is suspected of knowing about hacking, but I suppose there's a question as to whether it's worth hitting her with a corporate as well as individual charges. Les Hinton was CCed on the Goodman letter where he sprayed about the allegations about others' involvement in hacking (others who are now, it ought to be noted, being charged under RIPA). And Murdoch obviously signed off the Taylor settlement and then failed to notice any wrong when the Guardian splashed on it in 2009.

There's enough there to worry a few people. More generally, I suspect that a corporate conviction under RIPA might go down quite well with News International hacks who feel they have been thrown to the wolves to save the management (and one director in particular). It would also demonstrate that we are serious about corporate culpability. Worth keeping an eye on.

79 Criminal liability of directors etc.(1)Where an offence under any provision of this Act other than a provision of Part III is committed by a body corporate and is proved to have been committed with the consent or connivance of, or to be attributable to any neglect on the part of—
(a)a director, manager, secretary or other similar officer of the body corporate, or
(b)any person who was purporting to act in any such capacity,
he (as well as the body corporate) shall be guilty of that offence and liable to be proceeded against and punished accordingly.
(2)Where an offence under any provision of this Act other than a provision of Part III—
(a)is committed by a Scottish firm, and
(b)is proved to have been committed with the consent or connivance of, or to be attributable to any neglect on the part of, a partner of the firm,
he (as well as the firm) shall be guilty of that offence and liable to be proceeded against and punished accordingly.
(3)In this section “director”, in relation to a body corporate whose affairs are managed by its members, means a member of the body corporate.

Undead performance-related pay

The more I think about it, the more I think that a couple of recent developments in respect of performance-related reward for executives indicate a significant shift in opinion about motivational assumptions.

First up, if you read the Kay Review section dealing with pay it's pretty obvious that John Kay is a skeptic. He openly questions the value of paying bonuses, and draws attention to the way that performance-related reward might crowd out professional standards and/or attract the wrong kind of people. He seems to share the view that beyond a certain point, it really isn't about the money for many people. I don't mean that he thinks they don't care how much they are paid, but rather that tweaking reward systems isn't going to achieve a lot.

The Review's proposals - only pay extra reward in shares, and make executives hold them for a very long time - seem to me almost intended to do the opposite of the formal role of reward schemes. Kay is concerned about the potentially negative behavioural effects of short-term rewards, so he suggests we push the performance period right out. In other words, let's amend performance-related reward so that it has no immediate  effect.

You might think I'm overselling this, but at the Review launch last week, Kay was pretty strong on the way that we discount for the future. It really isn't much of a leap from this to the pay proposals.  If we defer the reward for 'performance' way into the future, it is unlikely to be valued much by the recipient and therefore won't exert much pull on executive behaviour.

In fact, if you accept my broader argument that performance-related pay is underpinned by a behaviourist psychological perspective then you ought to think that - from that viewpoint - pushing the reward way into the future is the wrong thing to do. Behaviourists argue that the reinforcer needs to be applied soon after the behaviour it is intended to reinforce. To state the obvious, if you only get your shares after you've left the company that isn't going to reinforce an action that you might have taken years earlier.

That's why I think that Kay's proposals may actually be a compromise designed to defuse the behavioural effects of performance-related reward within the existing system. I wonder whether Kay thinks such schemes are largely a waste of time, motivationally speaking, but has got the clear impression that there's not much appetite for scrapping existing practice. I have no direct evidence for this view, but otherwise the combination of the narrative about bonuses and the actual proposals on pay don't quite add up for me. And by the by, there's a great quote in Obliquity where he says carrots and sticks only work where you're employing donkeys and know exactly what you want your donkeys to do. I personally do not think that shareholder-focused remuneration systems can ever meet those requirements. I don't think John Kay does either.

In the same vein as Kay, Fidelity have come out with a proposal for "career shares" whereby directors would be required to hold at least some of their share-based rewards they receive until retirement.  Again, this must surely deaden any motivational effect of the awards. How can rewards of this nature have any greater motivational effect than providing a pension?

If, and it's still and if, we see executive reward develop in the way envisaged by Kay and others we will end up in a really odd place. We will continue to be providing performance-based pay, which has developed with the explicit aim of controlling/directing director behaviour, but it will be designed in a way that seeks to avoid any immediate behavioural effects. Performance-related reward will be in large part hollowed out - the schemes will still exist but without actually performing their supposed function.

In such an environment I think that "fairness" will start to become a bigger argument for performance-related reward (I've touched on this before). In fact Alfie Kohn made the point throughout Punished By Rewards that there is a moral argument buried in performance-based reward - if you do something well then you should get a reward, regardless of whether it motivates or reinforces. I suspect we're going to start seeing that line of argument supplant claims about motivation and/or alignment. But whilst companies and rec comms might use it as a way to defend performance-related incentives, arguing that executives deserve rewards might be quite a tough sell in an economy flat on its back with high unemployment.

Of course, all of this would be much easier to see, and easier to challenge if you disagree with it, if the motivational assumptions that sit behind various pay reform ideas were spelled out. But they very, very rarely are. As a result performance-related pay will likely continue to be widespread - and unthinkingly advocated - despite the fact that schemes aren't doing anything, in a motivational sense.

PS. Obviously I'm aware that some people attach an amazing "aligning" quality to the value of shares but it's not obvious to me why motivational effects should be any different to cash-based rewards.

TPA spin on unions

You may have noticed that the Taxpayers Alliance recently published its "name and shame" report on the pay of trade union general secretaries. What caught my eye was a piece in the Telegraph by the TPA's now chief executive Matthew Sinclair. It's a useful demonstration of their approach. For example:

In their latest returns to the Certification Officer, 36 trade unions have reported that they gave their general secretaries or chief executives pay and benefits of more than £100,000. That compares pretty well with the private sector. According to the Institute of Directors, the average basic pay of a managing director in a small company (turnover up to £5 million a year) was £70,000. At a medium-sized company (turnover up to £50 million a year) it was £100,000. And even in a large company (with a turnover up to £500 million a year) it was £128,000.

Let's initially just focus on the stats he quotes. There's a nice piece of sleight of hand in there. Many union leaders have "pay and benefits of more than £100,000" and this compares well to the private sector. Ok, so in order to compare presumably we'll look at pay and benefits in the private sector? Er... no. The comparison is actually with average basic pay. Guess what, if you and I are paid the same, and you compare my pay and benefits with your pay I will have a bigger figure. Matthew does compare one union leader's salary with the IoD figure, but further down in the article he only gives the total figures for three other union leaders. They clearly aren't the same type of figure as the IoD one. I suspect most readers won't notice the changes in figure used in the article, and will simply compare the two.

You might also question how valid the comparison is in any case. The TPA have looked at 36 union leaders, but there are more than 36 TUs. I assume all they have done is pull out the examples of those whose pay+benefits are over £100k - ie they've left out union leaders who don't fit the picture. They then compare those cases they seem to have selected on size with the IoD average figures. Comparing selected specific cases with averages, what could go wrong? Guess what - I could pick out some highly-paid IoD members who earn more than the average IoD member.


Finally, is the TPA even comparing like with like in terms of roles? The GS of a trade union is surely the equivalent of a chief executive or managing director of a business. If you're looking at an average of all directors, surely you need to compare with a comparable figure for unions - ie include at the least all DGS and AGS positions. If you can't do that then surely the TPA should compare union GS salaries with MD/chief executive salaries, not the stat for all directors.  


UPDATE: Matthew replies on Twitter that he is using average fig for managing director, so I was wrong about the last point. Not going to delete it as a) I should have spotted it (it's in the text) and b) I put up the incorrect claim so I should be accountable for it.

The Kay Review

It's got some great stuff in there. You can download the final report here.

Being a saddo, I find the most intersting stuff is about the nature of ideas that underpin existing market practice. I don't have time to do a proper overview so thought I'd bung up a few snippets I like.

Like this:

Regulatory philosophy influenced by the efficient market hypothesis has placed undue reliance on information disclosure as a response to divergences in knowledge and incentives across the equity investment chain.

 And this:

Anthropomorphisation of ‘the market’ in phrases such as ‘markets think’, or ‘the view of the market’ is common usage. It should hardly need saying that the market does not think, and that what is described as the view of ‘the market’ is simply some average of the views of market participants. ‘The market’ knows nothing except what market participants know

 And this:

measures to make the market more ‘efficient’, in the technical sense implied by the efficient market hypothesis, may have the effect of making the market less efficient in the broader and more important sense of achieving better resource allocation through better corporate decisions.

And the best line in the Review:

we see the sell-side analyst as a dispensable link in the chain of intermediation