Hacker News Reader

Comment by therealdrag0 | Hacker News Reader

> "Under a financial overhaul passed by Congress in 2010, the Securities and Exchange Commission was supposed to require all publicly held companies to disclose the ratio of C.E.O. pay to the median pay of all other employees, but it has so far failed to put it in effect. Corporate executives have vigorously opposed the idea, complaining it would be cumbersome and costly to implement."

-- Why would a ratio be costly to implement?

View on HN

Perhaps not costly, but definitely cumbersome.

What's more, it will have the opposite consequence. In his book "Influence", Robert Cialdini wrote that making CEO pay public had the effect of raising exec pay across the board rather than check it, as the lawmakers had hoped.

Isn't CEO pay of public companies already public? If so, the ratio would merely disclose the median employee's salary level.

I can't see how it would be that cumbersome, calculators are lightweight, widely available and easily fit in a pocket.

The thing about CEO pay rising after the information was made public was because it became a metric used by traders, so it started making financial sense to raise the CEO pay as it was thought to raise the company share price by more than the wage increase.

Of course this only works until someone matches your raise, then you have to raise again to keep the signaling going to the market and then the returns from this strategy rapidly diminish.

Now if the ratio was also public, then that would also become a market signal, but if you are a trader trying to gauge the health of a company then it should work in the favour of the employees, as if you are comparing two competing companies and trying to decide which one to bet on, if the CEO pay is the same and they are making similar profits, then the one with a smaller ratio will look like a much better bet.

edit - also, given that publicising CEO compensation can increase CEO compensation, why do you think that publicising employee compensation would reduce employee compensation?

Calculators are lightweight, but to decide which five digits to punch into the calculator, you might need to hire both an accountant and a HR lawyer.

Any company large enough to make this in anyway confusing has several of those on staff already.

Doing nothing? Most likely they are already busy with something, and adding this means more resources and cost to produce reports for the government.

If accounts and HR do not have the figures of what employees are paid already in a spreadsheet somewhere, in a form that is easy to run statistics on, then you have larger problems than producing reports for government.

And given you have not answered, why do you think that publicising employee compensation would reduce employee compensation?

You seem to be using the logic that the government tried something once and it did the opposite, so therefore all government policies do the opposite, rather than engaging with the potential outcomes of this specific idea.

As explained elsewhere, a figure of gross pay is definitely somewhere, but whether that is the figure required by law to be reported, and a figure that someone will think is what should honestly be reported, that's already another matter.

I did not argue that publicising employee compensation would reduce employee compensation, that was someone else. Perhaps you confuse authors.

yes, was confusing you with putlake, was tired, it began with a p and was about the same length, sorry.

Ah right. Indeed a very similar username. You sruley konw toshe mmees werhe txet can be raed if fsrit and lsat lrttees are rhgit.

I'm sure they can find some spare CPU power to tally up the CEO's earnings. I'll be glad to throw a Raspberry Pi in if it helps their cumbersome counting effort.

One way of interpreting that is that they think the disclosure would cause discontent that would lead to their labor expenses rising.

It isn't. Point me to a column of a table in a database containing all of your employee's salaries, tell me which one is the CEO's salary, and I can whip up a query to compute that ratio in less than a minute.

EDIT: Well, okay, if the given RDBMS doesn't contain rank or median functions, then it'll take a bit longer to implement. But not that much longer.

The costly part is that many companies don't have this available as simply as you assume.

As a BI consultant, this makes me feel very secure in my job and future business.

Because they dont want to reduce their own pay, which in turn would mean they have to pay their employees much more than they do now

Or they find other forms of compensation that do not equal "pay" to game the figures. I'm sure the army of lawyers out there can find a loophole or two.

If the CEo is also a chairman and major shareholder isn't this all meaningless twaddle

Is this information already publicly available? Can we calculate it ourselves?

It is most likely not the ratio part but the median pay of all other employees part. You need to keep an ordered list of the pay of every single other employee and keep it up to date as people join and leave the company. Then there is the issue of accounting for part time or temp workers.

And yet somehow they manage to do a payroll cycle every 2 weeks, produce W-2s every year and so on. I don't get it, obviously this would add up to quite a bit of admin work at a very large company but it's basically the sort of thing you can do very easily with a spreadsheet.

Assuming you know what data to include and what not. For instance, it is not obvious to me whether the "company" means a company in one country only, or all the legal entities of a multinational. Should you include part-time workers, and if you should how should you do it (the actual pay or the FTE equivalent pay, and what are the FTE hours for different jobs?)

It's a quagmire.

That doesn't sound like a quagmire, so much as a couple pages of typical regulatory detail.

That's what GAAP is for no?

If that's too much for someone to deal with they're probably not going to last long as a CEO. This is the kind of thing people have HR departments for, and you're telling me they can't crunch a few numbers? Come off it.

It's obviously possible to generate a figure; less obvious that the figure is actually meaningful, once people understand that subtle accounting differences like whether to include the salaries of the staff of the Manila office and whether to include total comp for the sales team with very aggressive bonus structures actually has more effect on the headline ratio than an above-market increase in CEO pay.

And perversely, laying off workers to outsource their jobs to an offshore BPO firm almost certainly gives the appearance the company is improving the staff pay...

It will definitely be meaningful specially for those huge multinational with CEO's earning extreme amount. Given that the article said something like 300:1 ratio of salaries in some companies even if your median wage salary is off by 2 i will still give you an idea that the CEO gets over 100 times the pay of a median worker. It might need some work to process and I am fairly certain the protest didn't start for that reason. Companies decided that they don't want to expose to massive gulp in payscale and then found the reason to protest.

It's definitely much _less_ meaningful for a multinational company than for one where all its workers live in the same city.

Certainly. Just to illustrate this, assume a small and simple multinational company that has headquarters and marketing (100 staff) in the US and manufacturing workforce (150 staff) in Vietnam. That's all the workforce.

The median salary in the company will be about $200 per month (because a majority of workforce is in a low-pay country). If the CEO makes $250,000 per year (which is not particularly much for a company of this size), she'll earn an outrageous 100 times as much as the median in the company.

Now let's assume that activists are unhappy with this 100 times difference and there's a boycott in order to reduce the CEO pay in proportion to workers. Guess what will happen?

Half of the workforce in Vietnam is kicked out and the gap in production is bought from a manufacturing partner in Pakistan (where the workers will earn $150 per month).

CEO pay is unchanged, but the median pay is now $2500 (pay of a not-high-status sales assistant in US). The CEO now earns only 10 times as much as the median worker!

Did things improve for the poor workers? Well, you could say so, because some people in Pakistan got a job, but the actual outcome still is that the people who work on the product now get less than before.

Road to hell is paved with good intentions, and naive activists are hit by the law of unintended consequences. Yes, the CEO-to-median figures are not very meaningful for a multinational company.

Let's be honest, and not pretend that this has absolutely anything to do with the difficulty of acquiring this information and providing it, and so much more to do with the "the less done to publicize how much I made, either by itself or in comparison to my employees, the better".

> You need to keep an ordered list of the pay of every single other employee and keep it up to date as people join and leave the company.

Is it even possible to run a company without having this information somewhere?

In most countries, certainly not. The tax office will require this information so it is already being stored somewhere.

However, "pay" is different from "wage". What you pay is quite straightforward. What is an equivalent wage or salary is not quite so clear.

Sure, there is a distinction between pay and wage. However, tax offices - at least here in the Netherlands, and the UK - are interested in equivalent wage, rather than the direct salary number.

Really? Tax office is not interested on what was actually paid, but a full-time equivalent pay?

I mean, let's take someone who is working three days per week (not unusual here) for a compensation of 1500 € per month. That means 60 % of full working time. He or she gets 60 % of "full time" salary. The tax office would want to know the FTE salary of 2500 € per month, and not what was actually paid?

I think we're not speaking of the same thing...

I think you're right, we're talking about different things :) I misinterpreted your comment - I was thinking of things like company cars, travel benefits etc. being considered part of the taxable pay package.

Ah yes - that is true as well. How to convert that kind of benefits into equivalent pay from which you calculate a median - that's not simple either.

For instance, in some country, a health insurance is tax-free and does not count towards a benefit, while in another it might be taxable income or considered part of pay. You could have expatriate employees in your team whose pay structure is entirely different from the rest. Now, if you have a multinational company, how do you put those employees into the table from which you start to calculate this figure for how much the CEO can be paid? Not simple.

I assure you that every company I've ever worked for has had that list of salaries at the tip of their fingers. This would be next to zero additional effort if it's only salary ratio. The problem is - you can easily play games with bonuses, stock options, etc... to make it appear that the CEO has a low salary.

So like

    SELECT MEDIAN(salary) FROM employees

(That's valid Oracle.)

I mean I get that there'd need to be some adjustments to normalize salary to yearly salary and factor in non-cash benefits, but come on.

So what you're saying is all that non-technical HR people have ever to do is normalize all of the employee salary data and load it into a database and then run this query? Easy peasy!

But in production this takes 3 Java developers, 2 Oracle "specialists", one year and a couple million dollars to implement.

Ah, back when I was a programmer without any SQL/DBA skills, I thought like this too. Now I know better than letting unskilled workers (e.g. lots of so-called DBAs, most developers) develop schemas or queries.

If you pay them all the median wage, that won't hurt convergence.

That was my thought as well. They need those numbers (or some contractor that handles payroll need those numbers) for reporting taxes etc anyway. There's no way that it is "cumbersome" to implement. I mean they need to total them up (salaries) for the quarterly reports -- and making one of those must be far more "cumbersome" than to report on a couple of core business stats...

The devil is in the details. Those "some adjustments" are not exactly trivial, if they need to be correct and undisputed. And different people may have different requirements for what is "correct".

Things get trickier when the company is multinational. You might even be prevented by law from handling the salary data of one country in another.

(FWIW, I have some years experience as a line manager in a multinational, with reports in another country; this was not always a breeze even though from what I know, we were a decade or two ahead of most multi-nationals in the level of international integration of corporate IT.)

If HR or Finance doesn't have this, they aren't doing their job. The company should know exactly how much money is going to payroll every pay period and should know exactly how much they are paying each employee. If they don't know this, the company has bigger problems.

What kind of a business doesn't know that already?

Now that I think of it, the best-run company I ever worked for was completely open-book; everybody could always see what everybody else was making.

Why yes, I run a company and have no idea how much my employees are compensated, and funnily enough, not even my own salary. This is probably why my books never balance every year. I knew shouldn't have fired our accountant.

> You need to keep an ordered list of the pay of every single other employee and keep it up to date as people join and leave the company.

Holy shit, you're right! We need a Hadoop cluster with AT LEAST 128 nodes to even come close to the amount of hardware necessary to process such data!

Reply on news.ycombinator.com