Unemployment explanations

When I knew noth­ing of eco­nomics, un­em­ploy­ment wasn’t mys­te­ri­ous. Peo­ple wanted a job, and couldn’t get one—well, peo­ple of­ten want stuff they can’t get. Noth­ing strange there, just one of those things.

Then I learnt some sim­ple eco­nomics, and it be­came more mys­te­ri­ous. The em­ploy­ment mar­ket is a mar­ket, with the salary be­ing the price. Why doesn’t this mar­ket clear? Why doesn’t the price (salary) sim­ply ad­just, and then ev­ery­one gets a job? It seemed profoundly mys­te­ri­ous that this didn’t hap­pen.

I’ve been grad­u­ally in­tro­duc­ing my­self to more eco­nomics (mostly in­di­rectly) and I’ve en­coun­tered a lot of ex­pla­na­tions for this per­pet­ual mar­ket failure. Thus the mys­tery of un­em­ploy­ment is, if not re­solved, at least some­what ex­plained. Since I would re­ally have en­joyed read­ing a col­lec­tion of un­em­ploy­ment ex­pla­na­tions when I was ini­tially puz­zled (al­most any ex­pla­na­tion of un­em­ploy­ment you read in the press is worth­less) I thought I’d do this for oth­ers. So here is my (en­tirely per­sonal and idiosyn­cratic) sum­mary of the main ex­pla­na­tions I’ve en­coun­tered.

Clas­si­cal ex­pla­na­tions: not a problem

The “re­vealed prefer­ences” idea holds that we can’t es­tab­lish what peo­ple want from what they say, but only from ob­serv­ing their ac­tions and their choices. So if peo­ple are not get­ting jobs, maybe they don’t re­ally want them? Or maybe they don’t want them, given gov­ern­ment un­em­ploy­ment in­surance?

I should note that what­ever you think of this ex­pla­na­tion, it ar­gues that un­em­ploy­ment is not an eco­nomic prob­lem. It views the un­em­ployed as hap­pily and ra­tio­nally choos­ing not to work, which thus suits their prefer­ences bet­ter and thus is of higher eco­nomic value (ne­glect­ing, of course, the cost of pro­vid­ing un­em­ploy­ment in­surance).

Clas­si­cal ex­pla­na­tions: mar­ket interference

If a mar­ket fails to clear, the first step is to look at out­side in­ter­ven­tions that may be pre­vent­ing it from do­ing so. Trade unions and gov­ern­ments are typ­i­cal can­di­dates for this. Trade unions en­gage in col­lec­tive bar­gain­ing, re­quiring that em­ploy­ers pay their work­ers a cer­tain min­i­mum level (in terms of salary, safety, or work­ing con­di­tions). This raises the price of em­ploy­ees above the mar­ket clear­ing level, and thus causes a short­age of jobs.

Similarly, gov­ern­ments can set min­i­mum wages, reg­u­late who is al­lowed to work in cer­tain pro­fes­sions, and tax em­ploy­ers or em­ploy­ees (the differ­ence is not rele­vant eco­nom­i­cally) for the job. They can im­pose other re­stric­tions, such as mak­ing it harder to fire em­ploy­ees, or cre­at­ing red tape that must be dealt with (thus rais­ing the cost of em­ploy­ing any­one).

Fric­tions and “Ir­ra­tional­ities”

The mar­ket equil­ibrium is an at­trac­tor state the mar­ket is sup­posed to move to­wards. Fric­tions are things that op­pose this move­ment or slow it down. “Ir­ra­tional­ities” are ways in which hu­mans differ from the perfectly in­formed self­ish ex­pected util­ity max­imisers of clas­si­cal eco­nomics. They may not be ir­ra­tional­ities in the stan­dard sense, just de­vi­a­tions from the model (note that un­like the stereo­type, economists are of­ten very aware of the limi­ta­tions of their mod­els).

There are costs in­volved in chang­ing most jobs. The em­ployee may need to change their so­cial cir­cles, their habits, their sched­ules, and may need to re­lo­cate to an­other area. The em­ployer may need to ar­range train­ing for a new worker, in­te­grate them into a team, and so on. Chang­ing a job is of­ten a very stress­ful event, im­ply­ing that em­ploy­ees, at least, feel these costs are sig­nifi­cant. In con­se­quence, peo­ple don’t re­lo­cate for a slight im­prove­ment of con­di­tions, and em­ploy­ers don’t fire peo­ple to re­place them with slightly cheaper work­ers. The mar­ket is thus less likely to ad­just.

Another fric­tion is the stick­i­ness of nom­i­nal wages. Peo­ple seem very un­will­ing to ac­cept a nom­i­nal pay cut, tak­ing this as an at­tack on their sta­tus. This pre­vents com­pa­nies from rapidly ad­just­ing to new eco­nomic con­di­tions (in both di­rec­tions: they are less likely to raise wages, if they can’t claw that raise back later) with their cur­rent em­ploy­ees. Some jobs have im­por­tant non-mon­e­tary as­pects to them (eg work­ing for gov­ern­ment and char­i­ties be­cause one wants to make a differ­ence, work­ing for a mar­ket leader, work­ing for a com­pany with a cer­tain cor­po­rate cul­ture, etc...); these as­pects (that form part of the whole com­pen­sa­tion pack­age) can­not be rapidly ad­justed up or down. Man­agers and work­ers of­ten de­velop re­la­tion­ships with each other, re­duc­ing the like­li­hood that bosses would rapidly fire work­ers or cut wages if eco­nomic con­di­tions de­manded it.

Hu­man be­havi­our and bi­ases can provide a gen­eral ex­pla­na­tion why agents fol­low be­havi­ours that make lit­tle eco­nomic sense (is­sues of sta­tus, of habit, prej­u­dices, etc...). Into this cat­e­gory go all the “rel­a­tive sta­tus” ex­pla­na­tions: it’s not im­por­tant to have a good stan­dard of liv­ing, but a com­par­a­tively good stan­dard of liv­ing.

Another in­ter­est­ing fric­tion is changes in the econ­omy pro­duced by new tech­nolo­gies, new in­no­va­tions, new com­pa­nies, and peo­ple en­ter­ing or leav­ing cer­tain in­dus­tries. The eco­nomic land­scape is always chang­ing, and all eco­nomic agents have difficulty ad­just­ing rapidly (be­cause of the fric­tions above, and be­cause of lack of in­for­ma­tion about the new setup, see next sec­tion). This ex­pla­na­tion is par­tic­u­larly in­ter­est­ing, be­cause they provide a pos­si­ble rea­sons that fric­tions wouldn’t die out, but could con­tinue for a long time. But if the equil­ibrium state is con­stantly chang­ing, then fric­tions can main­tain them­selves in­definitely, per­pet­u­ally slow­ing down any move to­ward labour mar­ket equil­ibrium.

In­for­ma­tion and agency

In­for­ma­tion eco­nomics is a fas­ci­nat­ing field, full of weird in­sights, in­clud­ing sev­eral that can con­tribute to un­em­ploy­ment. At the most ba­sic level, em­ploy­ees don’t know about the jobs on offer (they don’t know all job de­scrip­tions, and even those job de­scrip­tions only give a very par­tial im­pres­sion of what the job en­tails) and em­ploy­ers similarly don’t know about all po­ten­tial em­ploy­ees. There is a cost to find­ing this in­for­ma­tion.

In­for­ma­tion mar­kets are un­like any other mar­ket, though. Even the small­est of in­for­ma­tion er­rors can cre­ate new equil­ibriums—sta­ble equil­ibriums where the mar­ket doesn’t clear. What this means is that, lack of in­for­ma­tion can ex­plain sta­ble rates of in­vol­un­tary un­em­ploy­ment even in the ab­sence of any other fric­tions.

The prin­ci­pal agent prob­lem is an­other huge fac­tor here. If the “prin­ci­pal” re­quires an “agent” to do work for them, they can­not be sure the agent will do so in the way they ex­pect (this prob­lem can be seen as an in­for­ma­tion is­sue: the prin­ci­pal doesn’t know some­thing—the agent’s perfor­mance—and the agent can­not fully demon­strate it to them). Hiring man­agers may not fully trust their can­di­dates, and they might not be fully trusted by their bosses, who are not fully trusted by the share­hold­ers, and so on.

This means that worker X may be will­ing to do a job for salary S (in­clud­ing im­plicit re­mu­ner­a­tion), and firm Z may be will­ing to offer that, but nei­ther can trust the other to up­hold the deal. Ob­vi­ously this effect is stronger when the out­put can’t be mea­sured for­mally, or when im­plicit as­pects of the job (work­ing con­di­tions, office at­mo­sphere) are im­por­tant to the em­ployee.

This is one of the rea­sons that com­pa­nies of­ten pay “effi­ciency wages”, ie pay more than the mar­ket rate for jobs where out­put can’t be mea­sure for­mally. Em­ploy­ees who value be­ing part of the com­pany, have good re­la­tions with their bosses, and fear los­ing their good po­si­tion, will work harder to demon­strate their com­pe­tence, and bring more value to their em­ployer (in­tu­itive ex­pla­na­tion for this: the marginal worker, paid at the mar­ket rate, only prefers hav­ing a job to quit­ting it by an in­finites­i­mal amount, and will quite if any­thing goes slightly worse—is this the kind of em­ployee you’d like to have on staff?). Thus com­pa­nies that pay effi­ciency wages of­ten profit from do­ing so.

But pay­ing above the mar­ket rate, even for good rea­sons, still means that the mar­ket won’t clear. Note that effi­ciency wages alone are also enough to ex­plain in­vol­un­tary un­em­ploy­ment.


Macroe­co­nomics (by which I mean mainly the Key­ne­sian/​Mone­tarist schools) is built on the seem­ing triv­ial ob­ser­va­tion that the salaries of work­ers are the means by which they buy prod­ucts, and that the sale of these prod­ucts are the in­come of the firms that em­ploy peo­ple. There­fore, if the job mar­ket doesn’t “clear” it’s not as sim­ple as just wait­ing for salaries (the price of that mar­ket) to ad­just. If salaries fall, this feeds into less cash for con­sumers, which gen­er­ally im­plies less con­sump­tion, which feeds into less rev­enue for firms, low­er­ing their de­mand for work­ers, etc...

There are some the­o­rems that im­ply that the job mar­ket should still reach equil­ibrium… if there are no fric­tions. In the pres­ence of fric­tions, there is no rea­son to as­sume that the job mar­ket would clear: per­sis­tent in­vol­un­tary un­em­ploy­ment can be a per­ma­nent fea­ture of the econ­omy. In clas­si­cal eco­nomics, wages can­not rise in a sec­tor while a sin­gle per­son re­mains in­vol­un­tar­ily un­em­ployed (be­cause they could un­der­cut some­one and get their job or a frac­tion of their job); in prac­tice, wages do start ris­ing long be­fore in­vol­un­tary un­em­ploy­ment goes to zero (a “strong job mar­ket for pro­gram­mers” does not re­quire ev­ery as­piring pro­gram­mer to have a job). This raises the cost of hiring some­one, pre­vent­ing un­em­ploy­ment from dis­ap­pear­ing.

Fur­ther­more, the econ­omy need not set­tle down to a sta­ble equil­ibrium ei­ther: it can go through cy­cles of ex­pan­sion and con­trac­tion, thus ex­plain­ing the busi­ness cy­cle. This adds yet an­other ex­pla­na­tion for un­em­ploy­ment: the econ­omy be­ing in re­ces­sion.

The other claims of macro are that there are sev­eral differ­ent states that the econ­omy could be in, for the same given in­puts, and that fis­cal and mon­e­tary policy (gov­ern­ment tax­ing/​spend­ing and in­ter­est rate changes) can move it from one state to an­other, af­fect­ing the un­em­ploy­ment rate. The main prac­ti­cal differ­ences be­tween Key­ne­sian and Mone­tarists re­volve around the role of gov­ern­ment and what to do when in­ter­est rates are at zero, but that is not rele­vant for this anal­y­sis.

My evaluation

This is where I get to share my own un­var­nished and un­washed opinions as to the val­idity of these ar­gu­ments. Take this sec­tion with many caveats and doubts. Then add a few more.

First of all, clas­si­cal ex­pla­na­tions. Most of them seem off, for the same two rea­sons: be­ing un­em­ployed is very painful for most of the pop­u­la­tion, and most peo­ple care more about their rel­a­tive po­si­tion (are they as well off as their peers?) than about their ab­solute po­si­tion. This means that if some in­ter­ven­tion grad­u­ally re­duces ev­ery­one’s salaries by 50%, then most peo­ple will still be will­ing to work. Thus there seems to be lit­tle ex­plana­tory power in clas­si­cal ex­pla­na­tions that as­sume that peo­ple won’t work be­cause the salary is too low. This cat­e­gory also in­cludes em­ploy­ment taxes, since, who­ever nom­i­nally pays them, the ac­tual bur­den falls on those with the least flex­i­bil­ity, ie the em­ploy­ees (you can start talk­ing about elas­tic­ity here, if you want).

But firms are not in­di­vi­d­u­als, and func­tion very differ­ently. Thus clas­si­cal ex­pla­na­tions that ex­plain why firms won’t hire peo­ple (rather than why peo­ple won’t be hired) seem much stronger. This in­cludes min­i­mum wages and labour mar­ket rigidi­ties (eg firms can’t eas­ily fire peo­ple once hired). In prac­tice, min­i­mum wages seem to have lit­tle im­pact (there are a bunch of con­tra­dic­tory stud­ies here), but labour mar­ket rigidi­ties seem very im­por­tant. Cross coun­try com­par­i­sons (“South­ern Europe” un­em­ploy­ment rates ver­sus “North­ern Europe”/​An­glo Saxon ones) seem to bear this out.

I feel that labour mar­ket rigidi­ties along with the var­i­ous fric­tions and in­for­ma­tion is­sues (to a greater or lesser ex­tent) seem a good ex­pla­na­tion for the back­ground rate of un­em­ploy­ment of a coun­try (av­er­aged over the busi­ness cy­cle).

Con­versely, the Key­ne­sian/​Mone­tarist model seem good at ex­plain­ing the busi­ness cy­cle. Some of the clas­si­cal the­ory is also of value in ex­plain­ing re­ces­sions caused by ex­ter­nal shocks (eg the oil shocks of the 1970s).

Not explanations

No­tice what is ab­sent from these ex­pla­na­tions: gen­eral ed­u­ca­tion level, free trade (or its ab­sence), in­fras­truc­ture, tech­nol­ogy, com­pany X open­ing (or clos­ing) a fac­tory/​re­search cen­tre/​office in lo­ca­tion Y. It seems the sto­ries that are always re­ported in the me­dia may af­fect who gets a job (and at what wage), and may af­fect the over­all growth rate, but they don’t di­rectly af­fect the un­em­ploy­ment rate at all (in­di­rect effects may ex­ist, but they have to be analysed care­fully and are of­ten hard to pre­dict).

So. Now you (kinda) know.