Unemployed for six months: does the gap itself hurt you?

What the audit experiments actually measured

Three field experiments carry most of the weight in this area, and all three measured the same narrow thing: whether an employer responds to an application. Kroft, Lange and Notowidigdo submitted roughly 12,000 fictitious resumes to roughly 3,000 real online job postings in sales, customer service, administrative support and clerical categories across the 100 largest US metropolitan areas. The length of the current unemployment spell was randomly assigned between one and thirty-six months, which is what makes the result causal: nothing else about the applicant varied with the gap. Their finding was that the average callback rate declined sharply over the first eight months of unemployment and then stabilised. At eight months, callbacks were about 45 percent lower than at one month, with the rate falling from roughly 7 percent to roughly 4 percent. For scale, the authors compare this to Bertrand and Mullainathan's finding that resumes with Black-sounding names received about 33 percent fewer callbacks than those with White-sounding names. So the gap penalty is real and it is large. What none of these experiments measured is whether a person with your actual history, references, portfolio and network gets hired. They measured the first filter only, on invented people with no network at all.

The damage stops growing at around eight months

This is the finding almost nobody relays, and it is the one that matters most if you are seven months in. Kroft, Lange and Notowidigdo report that after eight months of unemployment, the marginal effect of each additional month of unemployment is negligible. The decline is front-loaded. Months nine through eighteen do not cost you what months one through eight cost you. Eriksson and Rooth, publishing in the American Economic Review in 2014, found a compatible pattern from a separate field experiment: employers attached a negative value to contemporary unemployment spells lasting at least nine months, but did not treat contemporary short-term spells differently, which the authors read as employers understanding that matching a worker to a firm takes time. Their second result is the more useful one for anyone reading this. Long-term unemployment spells in the past did not matter for employers' hiring decisions, which the authors interpret as subsequent work experience eliminating the negative signal. The practical translation is that the current gap is what gets priced. Once you are working again, this gap moves into your past, where the Swedish evidence says it stops being read as a signal. That is a reason to take an imperfect job, a contract, or part-time work sooner rather than holding out.

The gap hurts more where jobs are plentiful

Kroft, Lange and Notowidigdo tested how the callback penalty varied across metropolitan areas and found that duration dependence is significantly stronger when the local labour market is tighter. The result held across three different measures of tightness: metropolitan unemployment rates, vacancy-to-unemployment ratios, and the callback rate for a newly unemployed applicant in each area. This runs against intuition, and the mechanism the authors propose explains why. If employers use spell length as a signal about a candidate's unobserved quality, that signal is informative when jobs are easy to get, because staying unemployed for a year in a hot market suggests something. In a weak market, being unemployed says little, because everyone is. So employers discount the gap more heavily when hiring is otherwise easy. What you do with this: if you are searching in a strong local market and getting no responses, the gap is doing more work against you there than it would in a weaker one. That argues for routes that bypass resume screening entirely, such as referrals, contract-to-hire agencies and direct approaches to hiring managers, rather than sending more applications into the same filter.

You are not the person you think the numbers describe

Alan Krueger, Judd Cramer and David Cho examined who the long-term unemployed actually are for the Brookings Papers on Economic Activity in 2014. Comparing the long-term with the short-term unemployed, they found similarities in educational attainment other than among high school dropouts, similar distributions across industries and occupations, and few differences between urban and rural areas. Their conclusion was blunt: to a considerable extent, the long-term unemployed are an unlucky subset of the short-term unemployed. They also looked at the fourteen states with the lowest unemployment rates, where the average rate was 4.4 percent against 7.0 percent elsewhere, and found long-term unemployment in those states still reached 4.5 times its historical average during the recession. Being in a good local economy did not protect people. The point is not consolation. It is that the composition data does not support the story you are probably telling yourself, which is that people who are still unemployed at six months are the ones with something wrong with them. The differences the researchers found between the two groups were mostly demographic: a larger share of the long-term unemployed were over fifty and unmarried.

What happens after you get hired again

The same Brookings paper produced the number that argues hardest for taking the search seriously rather than taking any job and stopping. Krueger, Cramer and Cho found that only 11 percent of those who were long-term unemployed in a given month had returned to steady, full-time employment a year later. Reemployment did not reset the clock; a large share were jobless again soon after finding work. Two things follow. First, the first job back is a fragile object, and treating it as the end of the process is how people cycle. Keep the search apparatus running at low intensity for at least two quarters after you start: keep the network warm, keep applying to a small number of better-matched roles, keep the credential you started finishing. Second, if you have a choice between a role that puts you back in your own occupation and one that does not, the same paper found that long-term unemployed workers who returned to work tended to return to the same set of industries and occupations they had been displaced from. Sideways moves out of your field, taken in a panic, are the ones the data suggests are hardest to sustain.

Where a certification actually fits

Be precise about the mechanism, because this is where most advice goes vague. A certification does not shorten your unemployment spell, does not remove the dates from your resume, and no published field experiment has tested whether adding a credential to a long-gap resume raises callbacks. Anyone who tells you a certification cancels the gap is inventing a result. What a certification does is narrower and still useful. It converts an empty span of months into a dated, third-party-verifiable event that a screener can check in a registry. It gives you a specific answer to the interview question about what you have been doing, which is a question you will be asked and which you currently dread. And it can move you into a different applicant pool, where you are competing against people with less experience than you rather than against your former peer group. That last effect is the real one. If you are a laid-off operations manager applying to operations manager roles with a nine-month gap, the gap is compared against candidates who have none. If you are applying to cloud support roles with a foundational cloud credential, the comparison set changes. Choose the credential for the pool it moves you into, not for its reputation.

The five cheap credentials people reach for, and what each is for

CompTIA A+ is the standard help desk credential and requires passing two separate exams, Core 1 and Core 2, each needing its own voucher. Look hard at the destination before you buy it. The Bureau of Labor Statistics projects employment of computer support specialists to decline 3 percent between 2025 and 2035, with computer user support specialists declining 3 percent as organisations adopt automated troubleshooting tools. It still projects about 48,700 openings a year, all from people leaving the occupation, and the May 2025 median wage was $62,890. That is a shrinking occupation that still hires heavily, which is a specific situation, not a disaster and not a growth story. AWS Certified Cloud Practitioner costs $100 and AWS states it is designed for candidates new to cloud who may not have an IT background. Microsoft's Azure Fundamentals sits in the same slot; Microsoft does not publish a single price on its certification page, stating instead that price depends on the country or region where the exam is proctored, so check Pearson VUE for your location. CAPM from PMI requires a secondary school qualification plus at least 23 hours of project management education completed before the exam, and no work experience, which makes it one of the few project management credentials open to someone with a gap. ITIL 4 Foundation targets service desk and IT operations roles. Price it directly with PeopleCert or an accredited training organisation before committing, because this page could not verify a published exam fee for it.

Two things that make it worse

The first is fudging dates. Removing months and leaving only years, or stretching an end date to close a gap, is the most common response to six months out and it creates a specific failure mode: the discrepancy surfaces during background verification, after an offer, when you have already turned down other things. Employment dates are among the easiest facts to verify, because payroll records and prior-employer HR lines exist for exactly that purpose. A gap costs you callbacks. A date that does not match a payroll record costs you the offer and the reference. The second is optimising for software that may not work the way you have been told. You will read that applicant tracking systems automatically reject around 75 percent of resumes before a human sees them. No applicant tracking vendor, government agency or peer-reviewed study publishes that figure. It circulates without a traceable source, and it drives people to spend money on keyword-stuffing services instead of on the two things the research does support: getting into a different applicant pool, and getting a referral that skips the first filter. Format your resume so it parses cleanly, use the words that appear in the job advert, and then stop.

The next two weeks, in order

First, write down your actual last day worked. Not the month you started looking, the last day you were paid. Nine months from that date is the Eriksson and Rooth threshold, the point at which their employers began attaching a negative value to a contemporary spell, and eight months from it is where the Kroft, Lange and Notowidigdo decline flattens. Many people discover they are further from both than they assumed. Second, find any paid work that produces a payroll record, even part-time or contract, because both the Swedish evidence on past spells and the Brookings data on returning to work point the same way: the current spell is what gets priced, and ending it changes the object employers are reading. Third, pick at most one credential and pick it for the applicant pool it moves you into. Write down the exact job titles you will apply to with it and count how many are open within your commuting distance today. Under ten open postings means choose differently. Fourth, allocate your applications: if you are in a strong local labour market, the callback penalty is at its worst there, so weight your effort towards referrals and direct contact over volume applications. Fifth, set a date to stop studying and start applying with what you have, because the eight-month plateau means the cost of one more month of preparation is no longer rising, but the cost of not applying is.

What this does not fix

This page will not claim that a certification raises your callback rate. Every study cited here measured employer responses to fictitious resumes, and none of them tested adding a credential to a resume with a long gap. The research describes employer screening behaviour, which a certification does not directly change; what it changes is what you can put in the gap and which applicant pool you compete in.

Where an exam fits

• CompTIA A+, if you need a help desk role as a bridge and you accept that the occupation is projected to shrink while still hiring
• AWS Cloud Practitioner, if you are moving into cloud from a non-technical background and want the cheapest verifiable starting point at 100 dollars
• Azure Fundamentals (AZ-900), if the employers within your commuting distance advertise Azure rather than AWS
• ITIL 4 Foundation, if you are targeting service desk or IT operations roles where the job adverts name ITIL by title
• CAPM (Certified Associate in Project Management), if you have coordinated work before but never held the title, since CAPM requires no work experience

Common questions

Sources

1. 01Long-term unemployed (jobless 27 weeks or more) as a share of all unemployed people, July 2026 25.5 percent, or 1.8 million of 6.9 million unemployed
   US Bureau of Labor Statistics, The Employment Situation, July 2026 checked 2026-08-27
2. 02Median and mean duration of unemployment, seasonally adjusted, July 2026 median 10.5 weeks; mean 24.9 weeks
   US Bureau of Labor Statistics, Employment Situation Table A-12 checked 2026-08-27
3. 03Callback rate for fictitious resumes at eight months of unemployment compared with one month, in a randomised resume audit across the 100 largest US metropolitan areas about 45 percent lower, falling from roughly 7 percent to roughly 4 percent
   Kroft, Lange and Notowidigdo, Duration Dependence and Labor Market Conditions: Evidence from a Field Experiment (February 2013) checked 2026-08-27
4. 04Scale and design of the Kroft, Lange and Notowidigdo resume audit experiment roughly 12,000 resumes to roughly 3,000 job postings in 100 US cities, with unemployment spell randomly assigned from 1 to 36 months
   Kroft, Lange and Notowidigdo, Duration Dependence and Labor Market Conditions: Evidence from a Field Experiment (February 2013) checked 2026-08-27
5. 05Effect of additional months of unemployment beyond eight months on callback rates After eight months, the marginal effect of additional months of unemployment is negligible; the majority of the decline occurs in the first eight months
   Kroft, Lange and Notowidigdo, Duration Dependence and Labor Market Conditions: Evidence from a Field Experiment (February 2013) checked 2026-08-27
6. 06Whether duration dependence is stronger in tight or slack local labour markets, and the measures of tightness tested Significantly stronger when the local labour market is tighter, robust across metropolitan area unemployment rates, metropolitan area vacancy-unemployment ratios, and the callback rate for a newly unemployed individual in each MSA
   Kroft, Lange and Notowidigdo, Duration Dependence and Labor Market Conditions: Evidence from a Field Experiment (February 2013) checked 2026-08-27
7. 07Benchmark used by the same authors for the size of the gap penalty: callback gap for Black-sounding versus White-sounding names in Bertrand and Mullainathan (2004) about 33 percent fewer callbacks
   Kroft, Lange and Notowidigdo, Duration Dependence and Labor Market Conditions (February 2013), citing Bertrand and Mullainathan (2004) checked 2026-08-27
8. 08Effect of past versus current unemployment spells on employer hiring decisions in a field experiment past long-term spells did not matter; employers attached a negative value to current spells lasting at least nine months
   Eriksson and Rooth, Do Employers Use Unemployment as a Sorting Criterion When Hiring? Evidence from a Field Experiment, American Economic Review 104(3), March 2014 checked 2026-08-27
9. 09Share of people who were long-term unemployed in a given month who had returned to steady, full-time employment a year later 11 percent
   Krueger, Cramer and Cho, Are the Long-Term Unemployed on the Margins of the Labor Market?, Brookings Papers on Economic Activity, Spring 2014 checked 2026-08-27
10. 10Projected employment change and annual openings for computer support specialists, 2025 to 2035, with May 2025 median wage Overall decline of 3 percent, with computer user support specialists also declining 3 percent as organisations adopt automated troubleshooting; about 48,700 openings a year, all from replacement; median wage $62,890
    US Bureau of Labor Statistics, Occupational Outlook Handbook checked 2026-08-27
11. 11Cost of the AWS Certified Cloud Practitioner exam and its stated audience $100 (listed by AWS as 100 USD); designed for candidates new to cloud who may not have an IT background
    Amazon Web Services checked 2026-08-27
12. 12CAPM eligibility requirements a secondary school qualification plus at least 23 hours of project management education completed before the exam; no work experience required
    Project Management Institute checked 2026-08-27
13. 13Composition of the long-term unemployed compared with the short-term unemployed Similar in educational attainment other than high school dropouts, in industry and occupational distribution, and between urban and rural areas; the long-term unemployed are older and less likely to be married. The authors conclude that to a considerable extent they are an unlucky subset of the short-term unemployed.
    Krueger, Cramer and Cho, Are the Long-Term Unemployed on the Margins of the Labor Market?, Brookings Papers on Economic Activity, Spring 2014 checked 2026-08-28
14. 14Long-term unemployment in the fourteen US states with the lowest unemployment rates during the recession Average unemployment rate 4.4 percent in those states against 7.0 percent elsewhere, yet long-term unemployment there still reached 4.5 times its historical average
    Krueger, Cramer and Cho, Are the Long-Term Unemployed on the Margins of the Labor Market?, Brookings Papers on Economic Activity, Spring 2014 checked 2026-08-28
15. 15Number of exams required for CompTIA A+ certification Two: Core 1 covering hardware and networking, and Core 2 covering operating systems and security, which may be taken in either order
    CompTIA checked 2026-08-28
16. 16What Microsoft publishes about the price of the Azure Fundamentals (AZ-900) exam No single figure. The certification page states only that price is based on the country or region in which the exam is proctored, and directs candidates to Pearson VUE or Certiport to schedule.
    Microsoft Learn, Microsoft Certified: Azure Fundamentals checked 2026-08-28

Related