Scott Sumner on sticky wages for the unemployed

Scott writes (and Alex seconds, and here is a Woolsey response):

Nominal wages are fixed for the employed. NGDP falls 5%, and 5% of workers are laid off. Now the unemployed workers lower their wage demands by 20%. Why not by even more? Because of minimum wage laws, unemployment insurance, fear of loss of prestige, etc.

Suppose companies are not worried about workers making invidious comparisons (a big if, but I’ll grant this point to my opponents.) In the best case scenario firms lay off 4% percent of their workers and hire back the 5% who are unemployed at the same total wage bill. The excess unemployment is now 4% instead of 5%. The total unemployment rate falls from 10% to 9% (assuming 5% is the natural rate.) No big deal, we are still deep in recession. Thus wage flexibility among the unemployed doesn’t really help very much. If all employed workers accepted a 5% pay cut (or if the government ordered such a cut) and the Fed kept targeting inflation, we’d experience rapid economic growth.

Most of all, I praise Scott for being the only respondent, of many, to actually try and explain why nominal wages are sticky.

As the argument is presented, I would respond: “There is talk of fixed NGDP and talk of a fixed total wage bill.  Both conditions are assuming away the possibility of an easy solution.  An unemployed worker shows up, he and the employer cut a deal, a job is created, monetary velocity goes up, NGDP goes up, and the Pareto improvement occurs.  To use a different terminology, velocity should be elastic with respect to available gains from trade and thus so is NGDP.  More concretely, referring to the current day, employers are sitting on plenty of cash.  There is nothing in the NGDP identities to prevent further hiring from that stash.”

You can read Scott’s not-easy to-summarize counter-response to a comparable comment of mine here.  I do not interpret it as defending the relevance of nominal stickiness over the longer run or even as hoeing to the above passage.  For instance Scott writes “the actual change in NGDP is a sufficient statistic for understanding the net effect of wage flexibility on NGDP, PLUS monetary policy on NGDP.”  I believe Scott could have more accurately written”: “the actual change in NGDP is a sufficient statistic for understanding the net effect of labor market and other coordination breakdowns on NGDP, PLUS monetary policy on NGDP,” removing wage stickiness from the sentence altogether.  We’re back to viewing wage stickiness as one component of GDP problems, without knowing how important wage stickiness for the unemployed is, which is precisely where we started.


I will say what I said in comments to an earlier post...the research on job search suggests that reservation wages are not sticky and that the problem comes from the employer side. See the research citations to that comment.

wages are sticky because economies are generally characterized by longer expansions and shorter contractions in their business cycles. in the longer boom periods, a rise in output and 'productivity' improvements entail wage increases. this is the normal economic life. recessions are deviations from the normal economic trajectory, hence a fall in output is not followed by a decline in wages. when the economy starts to recover, the normal life-style continues.

"An unemployed worker shows up, he and the employer cut a deal"

What's about the worst thing one could say in a job interview? "I'm willing to work cheaper than the others." When an employer is judging a candidate they want a candidate that signals confidence not desperation. The "floor" wage level is always set by the employer not the employee. A potential employee can only ask for more money, not less, if she is expecting to be taken seriously.

Sure there may be some walks of life in which one-off transactions allow one to market themselves with cheaper service, such as lawn care ("I'll mow your lawn for $20.)" or prostitution ("I'll suck your dick for 10 bucks!"), but very few employers are in the business of bargain hunting for labor on an employee by employee basis. They control costs by firing unproductive workers not by hiring mediocre workers who might "offer value". One obvious evidence of the inflexibility of employers is this: try asking your boss if you can work half the time for half the pay and wait for the laughter to subside.

Wage stickiness is "employer side", as Donald mentions above. It is cultural and culture doesn't change easily. That's why when a company truly sets out to lower unit labor costs they do so by moving to a different culture.

I can attest to this (not the **** sucking!) from more normal economic times.

Over qualified does not mix well with a willingness to start out cheap and prove ones self. The employer that finally hired me told me so after a few months. The many, many who didn't I'm pretty sure were thinking the same way.

You pretty much refuted with experience what Rabbi Daniel Lapin suggested this past Sunday. He suggested that you should volunteer to work for free, and if you prove your worth, they'll pay to keep you before they lose you. I've seen this work well for internships where recent graduates have high abiliy but low experience, but ive never seen it work for a seasoned professional.

The real world is a lot harsher than some people would like it to be.

Per-employee costs beyond salary must be a big factor in this. Offering to work for $40,000 instead of $50,000 is insignificant against the fixed costs of health insurance, equipment, training, or the potential cost of firing a litigious employee.

The more complicated the job, or the more bureaucratic the system, the more the regular employees have essentially the same cost.

If the job is simple or commodity, you already expect to have low-cost interchangeable employees who can undercut each other. Likewise, in a free or liquid market, lower friction and overhead entails more flexible salaries.

If instead of a lower salary, you offered to not have health insurance, buy your own equipment, and not be a permanent employee, the employer will buy it if it is viable for the position. So you're a contractor or a vendor, which is very common. Voila.

"Offering to work for $40,000 instead of $50,000 is insignificant..."

I think this is right, but then we get back to lack of demand. What you are really taking is a slot they could give to someone else who might signal more potential.

But if being overqualified you can't underbid, then I wonder what those degrees are for.

If tax rates were higher on incomes, hiring workers would be a lot cheaper because the government would be paying a lot of the cost of hiring the worker.

Back in the 60s, Milton Friedman complained about all the doctors, dentists, and lawyers investing in wildcat oil wells which allowed hiring workers with the wealthy tax dodgers paying on $3 for every $7 the tax payer kicked in to drill the well. With one in ten wells coming up dry, the government was paying for 7 dry holes, the rich professionals 2 dry holes, and then the tax dodgers split the profits for the one producing well over the next few years or they got a capital gain taxed at only 35% by selling the well to a major oil company.

Today, the burden of failure falls mostly on the rich investor with taxpayers taking a small hit for failure, and if the investor strikes it rich, the taxpayers only get 15% for their limited risk. But investors risking their own money with taxpayers putting in such small contributions, $85 for every $15 from taxpayers, investors are looking for only sure things, like flipping real estate and betting on bad mortgages defaulting. And those don't involve hiring workers.

Back in the 60s, it was easier to write off losses on bad investments as well. Then you could hire an R&D team to fail and then quickly write off the losses and reduce taxes by 70% of the losses. In exchange for the lower tax rates, limits were placed on tax write-offs. The idea was to eliminate taxpayer subsidies for failure to make the economy for efficient with less failure.

That is perhaps the reason corporations want to increase immigration - by hiring the risk takers outside the US who have already reached a benchmark for success, corporations don't need to inefficiently pay people in the US to take risks.

If a corporation sells manufactured things, importing them pushes the risk taking of hiring outside the US where the governments do a lot to subsidize the risk taking that hiring people in the hopes they can efficiently deliver high quality products, instead of taking the risk that you can't find already expert people to immediately efficiently make high quality parts. Efficiency and quality is a matter risk taking over time to drive up quality and drive up efficiency which requires sustained investment in people, which drives up employment.

Today, the burden of failure falls mostly on the rich investor with taxpayers taking a small hit for failure, and if the investor strikes it rich, the taxpayers only get 15% for their limited risk.

In citing 15%, you're excluding the main form of conducting business, conduit entities, and ignoring all corporate taxes. Your premise rests a small sliver of the tax burden.

Second, if you make no capital contribution, exempt yourself from all capital calls, skim 35% of corporate profits every year without reinvesting or recontributing any, how can you say you've "invested"? That word doesn't seem to cover what you're describing.

If only aggregating wealth and spreading it lossily through bureaucrats and interest groups, were the same as paying an employee for his labor.

I'm not sure I understand the argument, but if I do it seems incorrect. If a deal is done to hire an extra worker, and this causes velocity to go up, the added velocity has just an infintesimal impact on the revenue earned by the firm hiring one more worker. Thus assume that hiring one more worker caused V to rise enough to increase NGDP by $100,000. Also assume the economy has 1 million identical firms. Then this $100,000 increase in NGDP will tend to increase the AD going to a typical firm by only 10 cents. Most of the gains from hiring the worker are "external", and hence the firm has no incentive to hire an additional worker, even if NGDP rose.

For the purpose of decisions about hiring, each firm can and should take aggregate NGDP as a given. It may be true that if all firms could coordiante, and agree to hire more workers, then V would increase enough to make all firms better off. But they can't.

Tyler also mentions the cash hoards that firms are sitting on, but that's not something that changes this argument at all. It's a given that firms maximize profits, whether they have big cash hoards or not. I don't dispute that certain actions by individual firms could boost hiring, and if all did this it might boost aggregate velocity by a lot, but those affects presumably have no bearing on the firm's decision-making. If only one firm behaves this way, the boost to aggregate velocity is too small to make the extra hiring profitable.

It seems to me that Tyler could just as well argue that (the Keynesian model implies) consumers should spend more, because if they did then NGDP would go up, and real income would go up, and jobs would increase, and hence there'd be enough income to justify the added expenditure. So I see his criticisim of sticky wage models as implying a similar criticism of sticky price models, indeed all of Keynesian economics, or perhaps I should say "Keynesian economics" because we are actually discussing the economics of Fisher, Hawtrey, Cassel, Pigou, etc.

BTW, there is an entire literature showing models where tiny menu costs can have huge aggregate effects. I wonder if Tyler is arguing that this literature is wrong, and if so, I wonder why he thinks it's wrong.

I'm also puzzled by the last paragraph. The problem isn't explaining NGDP, that's easy to do. The problem is explaining why low NGDP leads to low RGDP (if it indeed does.) Looking for ways in which wage flexibility might influence NGDP doesn't actually help solve the more difficult problem, which is why nominal shocks have real effects. I don't think I put as much emphasis on NGDP as Tyler assumes, at least in this case. I just find NGDP to be a convenient way of describing the size of the nominal shock that hit the economy. I can't even tell whether the use of "GDP"' in the final sentence of his post means RGDP or NGDP. It makes a huge difference which one he means. The fact that I can't tell indicates that I don't follow the argument he is making, and hence readers shouldn't put too much weight on this reply.

There are immediate private gains to the firm from hiring the worker for a lower wage; no doubt external gains may be bigger through a Blanchard-Kiyotaki effect, but I don't see where in your response you explain why the firm doesn't hire the worker. That is perhaps where we are stuck in this exchange. I don't see what stops the firm and I find that moving to the "NGDP level" only obscures the mechanisms at work. The gain to the firm may or may not be "infinitesimal" (though why should it be?) but if it is positive it should happen.

Why are we assuming marginal productivity is constant or at efficient levels?

Cant an employer lower the nominal wage by squeezing more work out of the retained work force? I dont necessarily mean more hours for the same pay, but that could certainly happen. I dont believe every worker is constantly at maximal performance. The threat of layoffs either disciplines the slackers or rewards the work-horses.

There's also substitution between high and low wage labor. A firm can keep its most productive workers and have then type their own memos and answer their own phones. They can also lay off their high wage workers and reprice their labor - getting younger people with lower salary expectations and a willingness to prove themselves at a bargain price.

Finally there is wage flexibility when jobs in higher wage sectors are lost and jobs in lower wage sectors are added. Some of the new low wage employees were once high wage employees. If the recovery is likely to be slow, then overqualification might not be a concern. Formerly high wage people are far more likely to accept a wage cut than you think. This effect is accentuated by the relatively low replacement rate of wages in unemployment insurance programs that are redistributive.

But i like the velocity argument presented her better than the last one. It uses macro concepts to describe macro phenomena. Youd have to convince me that the story of the employer and employee cutting a deal is scalable to the macro level.

I'm not sure that's true Dirk, wages just have to fall far enough. I think you are correct that at the margins hiring a lower skilled worker for a 10% pay cut isn't worth the productivity loss but hiring three guys at a 90% each just might be, especially in low skilled jobs. For example in my current environment I have a worker who makes about $80K a year whose sole job is to do a push a three separate buttons once a day each at different times. I would happily pay $8K a year for three workers with half his skillset to each push one of those buttons once a day. This scales across most industries. For example if wages wall far enough you could hire one sales guy per client, one doctor per patient, etc. You might take a hit in productivity but it would be offset immensely by the wages saved. Remember most workers (and even employers) aren't cream of the crop, they are happy with marginal barely productive workers.

What do you think is preventing you from hiring someone for say $50K to push those three buttons? I got to believe that you could easily find someone who would take that job for $50K if it is as simple as you suggest.

I've seen it happen. I watched guys in that exact situation strike themselves out of a job when mgmt was forced to figure out they could do 90% of the work without them. (Oops.)

The main reason it doesn't happen more often is just inertia, I think. Institutions tend to accumulate deadwood over time, for various and sundry reasons unrelated to the purposes of the shareholders/taxpayers.

You wouldn't rather pay a guy 80k to cover 10 clients than pay 10 guys 8k to cover 1 client each? The 80k guy will be more driven, gain experience faster, and much less likely to be offered a better opportunity at another firm.

We might say this is like sticky wages: The wages are sticky at small wage differences, but they shudder and snap to a far different wage if there is some structural change.

A few ideas for sticky wages:

Pay for Performance - Many firms say they pay for performance. It is advantageous for these firms to have only enough workers that all of their workers stay busy while performing to the top of their ability. Pay in this scenario needs to be proportional to each workers productivity which makes it difficult to reduce unit labor costs. This is a great incentive compared to arbitrary pay or pay not tied to performance.

Turnover Costs - Labor is not a commodity. Firms want the best workers with the most institutional knowledge working hard for them. Decreasing compensation sends a lot of workers away from your firm. While you may end up with the same "amount" of labor at less cost by restaffing, it often comes at the cost of having to retrain the workforce, lost sales while using inexperienced workers get up to speed and losing your best people.

Management Issues - Managers who have a choice between letting someone go or reducing pay for everyone would rather tell one person not to come in rather than have to deal with a whole team of discouraged workers every day.

Repeat after until learned.

Employers hire workers to produce goods and service for which there is a demand. The do not hire more workers than they need because the price of workers declines.

If an employer needs 4 workers it will hire 4 workers. If it can hire 5 workers for the price of 4, it will layoff the higher priced 4 workers and replace them with 4, not 5,of the lower priced workers.

Profits increase, aggregate demand decreases because of less consumer income, and the business may find it needs only 3 workers, so it lays off another worker.

This is basic Macro Econ 101. It is basic Keynesian economics. What is so hard to understand about this?

On the question is, "why don't labor markets clear?". I think you can see at least some of the reason by imagining a world where they did clear. In this world salaries would gyrate up and down with demand and supply similar to commodities. I don't think this would be the preferred system for either employers or employees. Employers would have to deal with widely fluctuating costs, making planning and forecasting of returns pretty hard. Yes it would help on downturns but probably a lot of the rent in upswings would have to be given to employees, who would likely be demanding higher wages to deal with insecurity (similar to the way returns on equity have to be higher than on debt). Employees would not be in favor either, given the choice between an average salary of $50k, consisting of equal years of $25k and $75k, versus a flat $50k, most people would elect to take the flat $50k. So to avoid this, employees and employers gravitate to flat nominal wages. In the US when you take a job there is an implied contract that nominal wages won't change much in bad times, as long as expectations in sharing of upside by employees is moderated. This is what people mean when they talk about morale issues with lowering wages in recessions, employees are pissed when they see the implicit contract broken, they got all the downside and none of the upside. Layoffs are easier to sell in this environment, there is no implied guarantee (at least in the US) of lifetime employment and they only affect a small group who are not around anyway to complain. As others have noted small layoffs can be raise efficiency as well, employers can get rid of troublesome or low value employees. So layoffs are the preferred method of adjusting costs for both employers and employees. The problem is that it creates sticky nominal wages. Because of money illusion, inflation is the only way employers can adjust wages without breaking their contract. Note that large layoffs with the intent of re-hiring at lower wages not only break the social contract but are just impractical, no organisation could continue to function given that level of disruption.

This only explains sticky wages for existing employees, why don't they also fall for new jobs where the contract has not yet been agreed? I think this must be partly due to reservation wages. For instance right now there is a lot of talk about robot drivers, with many people investing large sums into this technology. So at least some people think there is a demand for this. I would guess that 99% of unemployed Americans can drive so why are they not hiring themselves out as drivers to the market which is supposed anxious for robot drivers? Personal drivers are very normal in other parts of the world for the richer people, but you rarely hear of them in the US. Domestic servants are another area where Americans seem unwilling to accept jobs. I can imagine now the howls of outrage you told a university graduate they should get a job as a maid or driver and should forget that dream corporate job for now. Isn't that though an example of a reservation wage? In a recession, because of the implied contract to hold nominal wages steady, employees are getting rent, and unemployed people focus on trying to get a share of that rent rather than hiring themselves out for the more sure, but lower wage/status job. It is classic rent exhaustion.

The other reason for sticky wages for new jobs is that most new jobs require skills, and usually highly specialised skills. The labor market is not homogenous, and some areas are growing even when others are shrinking. An untrained person in a specialised job is literally a ZMP, so there is no market clearing rate that would allow an employer to hire them (example; an accountant trying to get a job as a doctor) . OTOH any potential employee with the needed skills in a growing area is likely to have other options, hence wages are not falling for this person. The only case where a new employee might be able to price himself into a job is a low skill area (where we are back to the reservation wage and most people won't take those jobs while there is at least a chance of getting a "proper job") or where the vacancy is in a shrinking area where there is a surplus of skills. But jobs in a non-growing area are likely to be pretty scarce, as those are the ones where the existing employees have the social contract and lay offs are, by definition happening in these areas. So wages don't fall much for new jobs either.

Oddly, your first paragraph is a pretty good description of certain segments of the tech job market over the past two decades. During the internet boom of the late 90s, people with minimal tech expertise could earn six-digit salaries, working at start-ups. This past decade, many of those then did the same work for half the salary at more staid firms. Whether that kind of salary gyration was preferred by employers or workers didn't much matter. Gyrate they did.

More, the job market is clearing in certain segments. There are many engineering specialties now where it isn't easy to hire, despite overall high unemployment. So I think your last paragraph is right. Neither jobs nor workers are interchangeable, so employment needs to be analyzed in terms of a segmented market.

I'm with David R on this. All my experience with layoffs is that employers view workers primarily in terms of getting work done. If there is less work to do (e.g. consulting circa 2003) workers are laid off. Similarly, if the company is trying to cut costs to maintain profitability, workers are laid off with the understanding somewhat less will get done (or that they're cutting deadwood).

So, why are wages sticky? I would say because employers don't care about wages nearly as much as they care about getting work done. They tend to view productivity as a personal trait and pay as something that depends much more on the type of work being performed and the type of person they're hiring than on the state of the labor market. Turnover hurts and labor market conditions change, so for most companies finding bargain labor isn't in their perceived best interests.

If all employed workers accepted a 5% pay cut (or if the government ordered such a cut) and the Fed kept targeting inflation, we’d experience rapid economic growth.

OK, I have to ask, because this reasoning boggled me the first time I saw it, and I'm still not clear how this argument is getting from A to B. Is the argument employers would hire more workers in this scenario? Or is this some kind of backdoor monetary stimulus in a way I'm not grokking?

If theyre laying off dead wood, average productivity is increasing. If every employee had formerly been earning the same wage, theyre getting more output per dollar of wages. Theyve got lower total output, but that's ok because demand is lower. We think too much of wages as dollar per hour than dollar per output. That is where we get wage flexibility.

But i dont believe productivity be worker is maximal and fixed. When employment is higher, manager monitoring costs are higher so average productivity is lower. After layoffs, monitoring costs per worker plunge, so productivity increases dramatically. This is why we see more output with fewer workers. Workers are not homogeneous and they do not have fixed qualities. If those were the cases, then let's just fire all the worthless HR application sifters and all that ineffective management who cant get people to do their jobs! Why are there so many books in the library on management and leadership if management cant effect worker and team output?

Policy regarding wage flexibility: it's pretty clear that you need sticky wages/prices to get Keynesian results. Lots of modeling effort has been devoted to justifying the assumption of sticky wages. But if flexible wages eliminate persistent unemployment, then why not shift policy towards more flexible wages? Policy could involve minimum wages, unemployment insurance, the length of labor contracts, etc.

Some will say that cutting wages reduces demand, but I think the formal modeling finds a real balance effect at work. Go through Patinkin on that subject.

Note re David R's comment: Aggregate demand certainly affects labor demand, but wage rate makes a difference. I walk into Starbucks; I see a line; I walk out. The manager thinks, "With an additional worker I'd have X fewer customers walk out after seeing how long the line is. How much would that worker cost me compared to the gross profits from a few more cups of overprice coffee?" In a simple Starbucks model there is a fixed relationship between workers and sales, but that's not how the real world works in many cases.

If wages were the main problem, money would be piling up with employees but it is mostly piling up at firms which indicates price stickiness. Rather than employees being overpaid, it is firms. Excess capacity prevents simply increasing investment, it must be new and different investment and believed to be profitable and scalable. Tech firms are doing some, but it mostly doesn't scale to large numbers of employees. A dirth of opportunities is seen, call it animal spirits, a lack of entrepreneurial zeal, technical stagnation beyond technology itself, or what, firms see more opportunity in buying each other than the hard work of creation.

Even the smallest companies can obtain flexibility by sending programming, data entry, customer service, etc out of state or out of country. Consultants pitch this on cold calls. You can expand and contract services daily. We'll see employment rise in India when growth commences.

Hah, when will that be O Wise One?

Metrodorus puts it best: “None of us knows anything, not even this, whether we know or we do not know; nor do we know what ‘to not know’ or ‘to know’ are, nor on the whole, whether anything is or is not”.

Let's say NGDP goes down 5% and wages stay at the same level for the unemployed. To make the math work, 90% of employees would stay employed at current wages and 10% would take a 50% wage cut. Or 9% would take a 55% wage cut. Or 5% would take a 100% wage cut. And so on.

More likely, 4% take a 100% wage cut and 5% take a 20% wage cut through being unemployed and taking a new job at a lower salary, adding up to 5% total cut from NGDP. The issue is that employers cannot easily offer one employee much less than all the other employees are making. The thinking goes that such an employee would jump soon "once the economy improves" or that such a wage would only get the most desperate employees. That may not show up in New Classical models, but it sure as hell shows up in spades in real life.

It is fairly clear that most economists have never actually hired, fired, laid-off, expanded or contracted a business.

Replacing the help with cheaper help requires a great deal of paperwork, training and orientation, etc. etc., which perhaps is a cause of the above mentioned inertia.
Time is a significant restraint for most managers, not to mention the "hassle factor." (I'm not certain if economists have a term for "hassle factor.")


All workers take a 5% pay cut. The firms all have lower costs. Each firm cuts prices a little to gain market share and expand profits more. All firms are doing this, so the price level falls. The Fed expands the quantity of money to get prices back up. Spending on output rises. Firms sell more. They produce more, and then hire more.

OK, so it's the backdoor monetary thing: create deflationary pressure and keep inflation targets constant.

Obviously firms can't expand market share in the aggregate since it always adds up to 100%, but prices should still fall in that scenario, so it's fairly plausible.

Or The Fed expands. Everybody expects prices to increase. Firms raise prices (and nominal wages). Production stays the same. No increase in hiring

Re: Unemployed workers taking a 5% paycut, does that include top income folks like corporate executives and the US Congress? Actually, it would be good if the salary for Congress were tied to the minimum wage and their health care plan were just a notch or two above what Medicade offers. Maybe then they would change their tune about a lot of things.

Hiring is done at the margin, and it is but one of the options available to an employer. Managers can use capital to replace labor through purchase of machines or software, use an outside vendor, or even go to another country to employ workers. Employers are not usually motivated to hire quickly. If they are in a production bind, where they can only meet supply by hiring workers, then they be. Otherwise, there is typically no benefit to rapidly increasing head count.

Salary may not accurately reflect the value of a worker to the employer. Intangible qualities (like the ability to fit in with the culture and to understand what's really important in a job) can make one worker far more valuable than another. However, unless an enterprise is exceptionally well run, management will not have much insight into the relative values of employees, and will tend to think that everyone in a given position is equally valuable. Managers tend to view employees as a fixed cost, with various overhead items (such as office space, payroll taxes, benefits, and IT support) that may far exceed a worker's salary. This makes employers somewhat insensitive to the salary portion of their cost of employment. It might take a very large salary differential to get management's attention. If salaries in India are 80% lower than salaries in the US, this might drive a change in strategy, but a 5% differential probably wouldn't.

Comments for this post are closed