Section 2

WFTC was introduced as anti-poverty strategy which increased inducements to work for low-income households with kids, with a peculiar concentration on relieving child poorness and inducements for lone parents, particularly individual female parents. It aimed to halve child poorness and raise employment of lone parents to 70 per centum by 2010. WFTC peculiarly marks individual female parents as they have a inclination to be hapless and face peculiarly high working costs13. From 1970-1990 employment in married female parents increased dramatically while the employment rate for individual female parents really fell to merely under 40 per centum, where it remained dead at 25 per centum below married female parents ; in 1970 the employment rates were approximately the same5. In comparing to other developed states, the UK ranked 15th out of the OECD states in 2001 in footings of per centum of individual female parents employed4, with a per centum employed half that of Portugal. This suggests there is the possible for a much higher proportion of individual female parents in the UK to be in the labour-force. These low employment rates have been responsible for the UK holding a high prevalence of child poorness with the greatest proportion of kids populating in workless families in OECD countries6. Harmonizing to the Institute for Fiscal Studies there are 2.9 million children1 life in poorness ( where family income is less than 60 per cent of the average national income, & A ; lb ; 236 a hebdomad ) . On the debut of WFTC, the authorities pledged to cut down this figure from 3.4 million in 1999 to 1.7 by 2010/20113. WFTC aims to cut down kid poorness through the redistribution of income to low-pay households, and besides incentivises those households to increase their income by traveling into work or increasing their hours.

In Section 1 we concentrated on the consequence on unemployment, nevertheless this is merely one portion of the strategy. In order to see the other effects of WFTC we should besides see the alteration in attempt induced, to see how the strategy changes incentives more by and large. Effort is a extremely of import factor as it influences the efficiency of the economic system. While public assistance strategies may see an betterment in unemployment rates, if the lessening in attempt outweighs the addition in attempt from lower unemployment so this is tantamount to a lessening in productiveness. Meghir and Phillips ( 2004 ) suggest that for low-skilled persons, hours worked is a good placeholder for effort13. This will besides assist to demo alterations in income.

Labour Supply and the Welfare Trap

Leisure is considered a normal good: when income additions, more will be consumed. As leisure additions, income must diminish so people will do the tradeoff between leisure and income in a manner that maximises their public-service corporation, dependant on their penchants. If the pay rate additions, two effects occur: the permutation consequence and the income consequence. The permutation consequence causes people to devour less leisure as it is now comparatively more expensive, while the income consequence causes people to devour more leisure as they can now afford more. Which of these two effects dominates depends on single penchants and the pay rate.

Welfare distribution strategies are frequently susceptible to the public assistance trap: participants in the strategy do non travel into employment or increase their hours as, due to the inducement construction set by the strategy, their fringy addition is non adequate to carry them to make so. For illustration, as people increase their work hours, the public assistance and other fringe benefits they receive lessenings, efficaciously doing a really high fringy revenue enhancement rate and therefore a important deterrence to work9.


The public assistance trap can be shown utilizing the Aid to Families with Dependent Children ( AFDC ) plan which existed in America from 1982 to 1996. AFDC created an income construction for a individual parents which is shown in figure X10, [ 1 ] . It allowed a public assistance grant ( G ) which remained changeless until a certain figure of hours ( h0 ) was reached. The grant would so be bit by bit reduced as hours increased until it reached nothing ( at h1 ) . During the decrease period, public assistance decreased as income increased on a one-to-one footing so income remained changeless ( Y1 ) , efficaciously making a 100 per centum fringy revenue enhancement rate. As a participant in the strategy there is no fiscal inducement to increase hours beyond this point. Eligibility of AFDC did non depend on hours worked and really decreased the inducement to be employed straight through an income consequence. The figure below besides shows the corresponding budget restraint and alterations in inducements for two individuals10,1. The original budget restraint is the line rudiment and the debut of AFDC causes the budget restraint to switch out to abdec, as the strategy participant will hold a higher disposable income for every hr worked.

The budget restraint diagram shows two persons ‘ penchants enduring from the public assistance trap. Individual A can increase their public-service corporation by increasing his leisure from LA1 to LA2 with no decrease of their income YA, shown by the displacement in the indifference curve from A1 to A2. B can increase their public-service corporation by really traveling into unemployment shown by the displacement in the indifference curve from B1 to B2, as they now have much higher income ( YB2 ) and can afford to devour more leisure.

This model created two typical results: those eligible for public assistance were improbable to work beyond h0 due to the 100 per centum fringy revenue enhancement rate and there was a distinguishable deficiency of benefits awarded to a individual parent who earns over the modest income of Y1, which was peculiarly low2,7 ( from above ) . This meant that there were negligible effects on both hours worked and poverty rates, which were reflected in the poorness figures in the US. Between 1988-93 merely 20 per centum of lone female parents who were on AFDC ( continuously for a twelvemonth ) were employed at any point and 50 per centum female-headed households were classed as life in poverty8.


The current US Earning Income Tax Credit ( EITC ) employs a different system and is a good illustration of how welfare strategies can make inducements to work. EITC uses a “ phase-in ” period, increasing public assistance as motion out of unemployment occurs and employs a more gradual “ phase-out ” period11. The “ phase-in ” period is designed to maximize the permutation consequence while understating the income consequence to supply greater inducements to get down work. The gradual “ phase-out ” period is implemented to better on the 100 percent revenue enhancement rate seen in AFDC. In figure X the debut of EITC causes the original budget restraint to switch out to the line abdec. The line Bachelor of Divinity represents the phase-out period and European Union, the phase-in period. The public assistance for person who does non work does non alter as their income remains at nothing. The permutation consequence of a pay addition is purely negative, but the income consequence is purely positive so these two effects work against each other to bring forth an equivocal result. Figure Ten shows two different sets of penchants for two persons which produce two different results in response to the debut of the EITC.

Individual A consumes leisure LA1 and receives income YA1 before the debut of EITC. When EITC is launched, A can increase their public assistance by working fewer hours and cut downing their income somewhat as shown by the displacement in indifference curves from A1 to A2. In this instance the income consequence outweighs the permutation consequence, demoing that EITC has caused a lessening in the figure of hours worked. However, for single B, the permutation consequence outweighs the income consequence doing both an addition in public assistance and hours worked. We can see from figure Ten that both the form of the new budget restraint and the penchants of the single determine the consequence of the public assistance strategy. Overall the consequence of such a strategy is equivocal depending on the figure of persons with certain penchants. The literature shows that EITC and subsequent enlargements had a positive consequence on engagement in the labor force12. EXPAND. Eissa and Hoynes ( 1998 ) found that EITC had the consequence of a lessening in hours for married work forces ( by 2 per centum ) and for married adult females ( by between 0.8 and 6 per centum ) , who are more likely to be at the higher terminal of hours worked, consistent with the theory. Elissa and Liebman ( 1996 ) studied the 1987 enlargement of EITC, which would hold had much the same consequence dependant on hours as shown above, and found an undistinguished consequence on those already working.


Where the EITC ‘s cardinal purpose was to diminish public assistance payments, FC and WFTC ‘s purpose is to diminish poorness through increased public assistance and inducements to work12. FC used a somewhat different construction utilizing a needed figure of 16 hours for eligibility, greatly increasing inducements to make this degree of work. There was besides an extra child care recognition for those working above 30 hours, making an excess inducement to make this mark. Diagrammatically, it creates a “ corner ” on the budget restraint which is more likely to capture more indifference curves. As in EITC, FC still used the “ phase-out ” period to halt the 100 percent fringy rate of revenue enhancement. WFTC attempts to better further on FC, as explained in subdivision 1, by being more generous: it increased recognition for kids under 10, had a higher income threshold, increased public assistance for child care costs and had lower taper rate in the phase-out period ( diminishing from 70 to 55 per centum ) . A comparing of the two strategies ‘ constructions is shown below in figure X9.

The increased generousness of WTFC strategy changed inducements for people with kids depending on their hours worked. The first group are those who are unemployed: WFTC increases an already unequivocally positive inducement to work. From this group we would anticipate additions in employment. The 2nd group are those who were working below the FC eligible income threshold. For this group, if the person was already working at the corner solution of 30 hours, it is likely that they will be at the same 30 hr corner solution post-reform. This is because they will lose extra child care recognition if they decrease their hours and are improbable to increase hours as the fringy revenue enhancement rate is merely decreased somewhat. Outside of this corner solution WFTC will hold a assorted consequence depending on the magnitude of the income and permutation effects.

The 3rd group are low-income workers who used to be ineligible but due to the increased income threshold are now eligible. WFTC has an unequivocally negative consequence on this group as it increases income and significantly increases their fringy revenue enhancement rate. Specifically for the WFTC, it increases their old fringy revenue enhancement rate from 33 per centum to merely under 70 percent9. The 4th group are those who work merely above the new income threshold who may be incentivised to take down their hours and choose in to the public assistance system ; WFTC here creates inducements to diminish hours worked. One of import note to do is that the lower the taper rate, the lower the fringy revenue enhancement when on public assistance, but the greater the figure of people it might impact negatively, who have a comparatively high income already. The overall consequence on hours worked of such a alteration in the public assistance strategy hence depend on the comparative sizes of the groups and the magnitude of the behavioral responses of the groups10. The consequence on these groups are shown below in figure X. The gradient alteration of the budget restraint reflects the reduced taper rate.

Consequence on individual parents and twosomes

We can utilize the effects above to find the likely consequence on married twosomes and solitary parents who have kids. For unemployed individual parents, as discussed above, WFTC will hold an unequivocally positive consequence. For working individual parents, the consequence will be really dependent on their current hours worked and penchants as explained above and the terminal consequence on labor will be equivocal. Therefore there should be an addition in labour-force engagement and an equivocal consequence on hours worked for.

If we consider a married twosome with a individual earner, there will be really much the same inducements as a working individual parent where inducements depend on current hours worked and penchants ; the consequence on hours worked will be equivocal although Blundel et Al ( 1987 ) suggested that the addition in the engagement rate is improbable to be dominated by lessenings in hours worked10. Concentrating now on married twosomes who are both unemployed, as in the individual parents instance, there is an unequivocally positive inducement for one of the brace to travel into employment and we should anticipate higher engagement in the labour force. If we consider a married twosome where there is one chief earner, the household income has now increased doing a big income consequence for the secondary earner, which creates an inducement to diminish hours or to travel out of work wholly. Blundel ( 2000 ) showed that WFTC created both a positive income consequence and a negative permutation consequence, making a dual inducement to cut down hours, for those working under a certain figure of hours, dependant on the chief earner ‘s income [ 2 ] . The demand for child care costs of both twosomes working over 16 hours could besides ensue in an addition in hours worked10. From married twosomes with one chief working spouse, we would anticipate a lessening in employment rates for the secondary earner and equivocal effects on hours worked for the chief earner. Need TO CHECK THE ELASTICITIES ARE WHAT YOU THINK.

Most of the literature suggests that males tend to be extremely unresponsive to alterations in benefit inducements with respects to working hours with some pay snaps estimated every bit low as 0.0614 and -0.0415. On the other manus, male engagement rates seem to be influenced to a great extent by public assistance strategies such as WFTC13. Females ‘ engagement rates and working hours tend to be sensitive to welfare inducements, particularly for lone female parents, where engagement snap for lone female parents with regard to in-work benefits has been measured at up to 1.816 and Brewer et Al ( 2005 ) found it is one of the highest of the demographic groups. These snaps give a good thought of the effects of the public assistance and which demographic is most sensitive to welfare strategies, nevertheless, they do non give the full image. Due the non-convex form of the income construction created by a public assistance strategy such at WFTC, negative inducements to work are besides created13 ( as seen in groups 3 and 4 in figure Ten ) .

Section 3

In Section 1 we used the Family Resources Survey which took a random sample of adult females at monthly clip periods. We so compared the proportion of lone female parents who were employed before the execution of WFTC, with the proportion of lone female parents who were employed after WFTC, utilizing a control group to account for tendencies and unexpected dazes. This information basically allowed us to compare the difference in the proportion employed between one random sample collected pre-reform and another random sample post-reform, instead than tracking the alterations that occurred in the pre-reform random sample. This could ensue in colored consequences if there was a alteration in composing and the samples of before and after WFTC differed in features which were non included in the theoretical account. To acquire more accurate estimations we can utilize panel informations which would give perennial observations on the same persons over certain clip periods.

Panel information allows us insight into the kineticss of the informations, which cross-sectional informations can non supply. For illustration, it helps us find whether additions in engagement are due to an addition in those fall ining the work force or fewer go forthing it. Panel information besides gives us the chance to command for underlying unseen differences in the persons ( unseen heterogeneousness ) which would otherwise cause omitted variable prejudice. For illustration, in our WFTC sample, there may be an implicit in leaning to work or unseen ability which differs for each person and is non easy included in the theoretical account. This omitted variable prejudice could in theory be corrected utilizing instrumental variables when utilizing cross-sectional informations, but it is frequently hard to happen suited instruments. The usage of panel informations will besides give us more efficient estimations for the extra ground that there tends to be a higher figure of grades of freedom than with transverse sectional informations.

I will concentrate on the WFTC ‘s consequence on lone female parents because WFTC is peculiarly aimed at this demographic and work forces tend to work full clip and have a low response to additions in income13. I will hence utilize the same control and intervention groups as used with the FRS informations: the intervention group will dwell of lone female parents while the control group will dwell of individual adult females without kids as they are ineligible for WFTC.

Panel Data

The information I will utilize comes from the British Household Survey ( BHPS ) which follows the sample persons twelvemonth after twelvemonth. In line with Francesconi and Van der Klauuw ( 2006 ) , the informations includes individual females who are over 16 and are at a maximal age of 60 in any twelvemonth over the period 1991-2003. The information is limited to before 2003, as there was farther reform in 2003 when WFTC was separated into Working Tax Credit and Child Tax Credit. I besides included annually GDP informations to account for the effects of the concern rhythm on unemployment. As in subdivision 1, I will take inconsistent informations such as those claiming to be a childless females with kids or in paid work while working zero hours, and take entries with losing informations necessary for our theoretical account. This consequences in a pooled sample of 7,948 observations including 1,995 alone individual adult females, 1,499 of which are individual females without kids and 756 are solitary parents. Table Ten shows drumhead statistics of the informations used. From the tabular array we can see that lone female parents are much more likely to work over 16 hours and over 30 hours as they have to take clip to care for their kids. The chance of remaining in a occupation ( employment continuity ) and the chance of come ining a occupation in a given twelvemonth is loosely similar for the two groups over the period, as is the mean age and age go forthing full clip instruction. Single adult females on the other manus, earn on mean 110 % more a month than lone female parents.

We can see there was a sudden upward tendency in the proportion of lone parents employed in 1997 which continued to the terminal of the information ; this is besides the instance for hours worked as the two are interlinked. How much of this station 1999 addition in both employment and hours can be attributed to WFTC is problematic.

The undermentioned figures plot the proportions of those who enter work in a given twelvemonth from being unemployed, or remain in a occupation for back-to-back old ages. The figures show a volatile but diminishing tendency for the proportion come ining the work force for individual adult females without kids combined with a somewhat increasing tendency of those remaining in the work force. This suggests that the little lessening in employment from 1991 for individual adult females without kids is due to a lessening in people come ining work. Lone parents have a volatile and really somewhat increasing tendency of people come ining work combined with a steeper upward tendency for those remaining in work. This would connote that the addition in continuity has been responsible for most of the addition in employment for lone female parents.

The Model

There are a figure of theoretical accounts which can be used to gauge effects when utilizing panel informations. The pooled OLS theoretical account combines the information over I and T into a arrested development with NT observations to give changeless coefficients of the theoretical account:

yit = a + bxit + uit I = 1, 2… N t = 1,2… Thymine

Where yit is the dependent variable, xit is the explanatory variable and uit is the disturbance term. For consistence, the theoretical account requires the error term to be uncorrelated with the explanatory variables both over I and T for consistence. Due to unseen heterogeneousness, the mistake footings are likely to be correlated for an person over the clip period and the calculator will be inconsistent. We must therefore use an individual-specific effects theoretical account as shown below:

yit = ai + bxit + uit I = 1, 2… N t = 1,2… Thymine

Where Army Intelligence are random variables which differ between persons but are changeless over an person ‘s clip period: it allows persons different intercept footings, which are changeless in every clip period and therefore capture the unseen heterogeneousness. The explanatory variable is xit and uit is the mistake term which is assumed to be independently and identically distributed. Taking the norm over the clip period we get: I = 1, 2… N

Subtracting the two we find: I = 1, 2… N t = 1,2… Thymine

If we estimate this equation by OLS it will give us the same estimation of B as the original theoretical account, which is the fringy consequence of xit on yit. This is the fixed effects ( FE ) theoretical account. The single feature is equal to its norm over the clip period, as we have assumed it is changeless over clip. This means that when we take the difference it does non have in the concluding equation. However, this technique does non let for appraisal of the effects of time-invariant explanatory variables as they would be absolutely collinear with the single effects. In this illustration it would intend that would be 0 and therefore B could non be estimated.

In order to gauge this equation by OLS, the Linear Probability Model ( LPM ) will hold to be used as our employment dependant variable is binary, that is person is either employed or they are non. If written in footings of divergences from 0 or 1 and the chances of the divergence occurring, the outlook of the mistake footings in the theoretical account I = 1, 2… N

can be written as Calculating the discrepancy of the error term ( E [ ui2 ] ) we get and therefore the random error term suffers from heteroskedasticity. In order to counterbalance for this we must utilize the divide our theoretical account by the weights which will do the mistake term homoskedastic. This theoretical account can be used in concurrence with the FE theoretical account. For simpleness I will non utilize the Logit theoretical account which is merely purely needed if the LPM predicts chances above 1 or below 0, which does non happen in the consequences.

The random effects ( RE ) theoretical account is an alternate individual-specific effects theoretical account which requires the single features to be distributed indiscriminately among persons but still constant over clip. If this is the instance so the RE theoretical account is consistent and asymptotically efficient, utilizing the discrepancy within and between persons and GLS or FGLS to make efficient weights, whereas the FE theoretical account is merely consistent, as it merely allows for fluctuation within the person. However, if the fixed effects are non indiscriminately distributed but are correlated with the explanatory variables so the RE theoretical account is inconsistent and the FE theoretical account should be used. The RE theoretical account besides allows us to include and gauge the effects of time-invariant explanatory variables. We can prove the necessity for the FE theoretical account utilizing the Hausman trial for endogeneity. The Hausman trial can be used to set up whether the difference between the fixed effects and random effects calculators is statistically important. The void hypothesis of the trial is that there is no correlativity between the single effects and the regressors and the RE theoretical account can be used. In order to corroborate our demand for the FE theoretical account, we would necessitate a big Hausman trial statistic and a important P value. It is likely that the composing of the individual adult females group will hold changed, as WFTC has the consequence of doing being individual more financially attractive, and therefore the FE theoretical account should be used15.

As panel informations includes cross-sectional and clip series informations, there is the potency for both consecutive correlativity and heteroskedasticity. The FE theoretical account reduces consecutive correlativity significantly as it captures time-invariant unobservable features but there may be heteroskedasticity nowadays in the information. Although this will non consequence the estimations, it will impact the standard mistakes, so I will utilize robust criterion mistakes in the arrested development to account for this.

The difference-in-differences method ( DD ) requires the intervention group and the control group to portion the same tendencies and dazes to give accurate consequences as discussed antecedently. However, if we look at figure X, visually it appears that the two groups do non portion a common tendency as employment for lone parents is increasing significantly more than individual adult females without kids before the proclamation and execution of WFTC. If we use the difference in difference method, we could wrongly impute this already increasing employment tendency to WFTC, annuling our analysis. As in Francesconi and van der Klaauw ( 2006 ) I will run another arrested development utilizing the difference-in-difference-in-difference ( DDD ) method as a comparing which includes a clip variable and a clip variable interacting with the intervention group, leting for different tendencies between the groups.

Therefore I will utilize the undermentioned theoretical account which includes the DDD method: I = 1, 2… N t = 1,2… T Where the theoretical account result variable yit is employment, dot is a dummy variable for being a lone parent or non, allows for tendencies and besides different tendencies between the two groups ( the DDD method ) , platinum is the silent person variable for being station reform, is a K?1 vector of variables which could impact employment, qi are the fixed effects and uit is the independently and identically distributed error term.

As the BHPS information is study informations, there are classs which are provinces of being and are non quantifiable. For illustration we can non find the numerical difference between life in London and populating in Scotland. The person will either unrecorded in London or will non. In this manner we need to make a figure of dummy variables for each of the options in the unquantifiable classs in order to demo the consequence, for illustration, of life in London compared with non populating in London. We besides have to except one of the options in each class in our theoretical account when we apply the silent person variables as the theoretical account will happen the effects of the other options in comparing to the excluded option. From the arrested development we get the consequences:

These consequences show the impact of WFTC on different types of employment utilizing a figure of theoretical accounts. The coefficients in table X show the fringy consequence of the variable on the proportion of people employed. The fixed effects difference-in-difference ( DD ) theoretical account shows WFTC is responsible for an 8.2 per centum addition in employment, which is important at the 5 per centum degree: the coefficient is significantly different from nothing with over a 95 per centum assurance degree. This positive coefficient is consistent with the theory, where WFTC unequivocally increases inducements to work for the unemployed. The other statistically important coefficients behave as expected, for illustration holding a first grade increases the likeliness that the person will be employed while holding immature kids reduces the likeliness. The consequences show that the younger the kid, the more clip will be needed to care for them and so the fewer the hours worked. House ownership has a high criterion mistake and is non important in all the theoretical accounts, most likely because people will either work full clip and have adequate money to purchase a house, or will hold adequate money to have a house and besides be retired. As the OLS pooled theoretical account does non command for fixed effects it is likely that there will be consecutive correlativity which by and large leads to underestimating coefficients and overrating t-statistics16. This over-estimating of t-test statistics can be seen in the consequences as the OLS pooled theoretical account estimations that 43 coefficients are important at the 10 per centum degree or better compared to the FE theoretical account which finds 28 coefficients statistically important.

We can prove the cogency of the fixed effects theoretical account here utilizing the Hausman trial. The trial produces a big trial statistic ( 162.14 ) and a p-value of 0.000. This leads us to believe we should reject the void hypothesis that there is no correlativity between the single effects and the regressors and corroborate our demand for the FE theoretical account.

The FE theoretical account which uses difference-in-difference-in-difference includes a clip tendency and a clip tendency interacting with the lone female parents. The clip coefficient suggests that the sample as a whole has a diminishing employment tendency. However, the positive coefficient for the clip tendency interacting with lone female parents suggests that lone female parents are sing a different tendency, with employment lifting by 1.2 per centum a twelvemonth. This would propose that the tendencies between the two groups are so different and DDD is necessary to happen the true WFTC consequence. Compared to the DD theoretical account, the estimated consequence is much smaller at 2.8 per centum as the already increasing tendency in employment for lone female parents relative to the tendency for childless individual adult females is no longer incorporated. However, statistical significance is besides lost as clip variables will be correlated to some extent and it is harder to place which clip variable is the cause for increasing employment. Francesconi and van der Klauuw ( 2006 ) happen a statistically important consequence of 5.8 per centum on employment when utilizing a FE ( DDD ) theoretical account. It is likely that this consequence is greater due to the difference in explanatory variables, despite utilizing the same general theoretical account, and the sample size will hold affected the significance: my figure of observations is much fewer.

The theory suggests that for those who are unemployed or work under 16 hours, there is an unambiguous inducement to increase hours to 16, where eligibility for WFTC occurs. The consequences confirm this, demoing that WFTC has the consequence of increasing the figure of people who work 16 hours or over by 9.6 per centum. This figure is greater than the consequence on labour-force engagement, because it increases inducements for both the unemployed and those working under 16 hours, to work 16 hours and over. Francesconi and van der Klauuw ( 2006 ) find that the consequence on labour force engagement is greater than that of the consequence on eligible employment which seems unusual as it implies that WFTC encouraged some antecedently unemployed lone female parents to work without really having any of the in-work credits. If they did have in-work credits so they should be counted in the consequence on eligible employment.

From the consequences, we can see that WFTC ‘s consequence on full-time employment is non significantly different to 0, proposing that the reform did non make an inducement construction which encouraged people to travel to full clip employment from unemployment or to increase their on the job clip to over 30 hours. The theory in this instance shows that it is improbable that person would work merely merely below ( merely 1.2 per centum of the sample work 27-29 hours ) . The non-negative consequence besides suggests that people were non incentivised to diminish their on the job clip below 30 hours as they face the big deterrence of losing their extra child care recognition [ a ] .

The theory predicts that WFTC will hold an consequence on the figure of hours that individual female parents work, dependent on the figure of hours they worked before the reform. The above consequences can non corroborate this anticipation as they merely show motions between unemployment, eligible employment and full-time employment instead than motions within these types of employment.


The grounds from the informations supports the theory that WFTC unequivocally increases inducements to come in work. However, the theory concludes that the overall consequence of WFTC on hours is equivocal. In order to happen the terminal consequence of WFTC on overall hours we can utilize the panel informations and the same theoretical account as before [ B ] but utilizing the log of occupation hours as the dependant variable in order to accomplish per centum alterations in hours instead than absolute values.

While the fixed consequence theoretical account allows for single features, the coefficients will be inconsistent as we will be utilizing censored informations. This censorship occurs because our dependent variable can non be observed at all provinces: if harmonizing to the theoretical account the person ‘s penchants deem that they should work negative hours, they are capable to a lower bound and can merely take to work a lower limit of nothing hours and be unemployed. The true ( negative ) value in this instance is hence non observed.

The consequences show the consequence of WFTC on persons working different hours utilizing a figure of theoretical accounts. The FE ( DD ) arrested development for entire occupation hours shows that the entire figure of hours worked for lone parents increased by 3.01 per centum as a consequence of WFTC. Again, as in the employment arrested development, the smaller WFTC consequence coefficient when utilizing the FE ( DDD ) arrested development implies that some of the upward tendency in hours ( seen in figure X ( average hours graph ) ) is wrongly attributed to WFTC and significance is lost as designation is harder among more clip variables.

As our informations is censored, the deduction is that the FE theoretical accounts ‘ coefficients are inconsistent and so the Tobit theoretical account is used as a comparing. I will utilize the same theoretical account as in subdivision 1 [ B ] with the dependant variable as the natural log of the hours worked and utilizing a Tobit arrested development which requires:

The Tobit theoretical account coefficients can so be used to cipher the fringy effects [ hundred ] . From the consequences, these fringy effects appear to be in line with the FE ( DD ) theoretical account and are statistically important at the 5 per centum degree. The “ in work ” arrested development shows how entire hours changed for the on the job population, taking the impact of increased employment on hours. The positive statistically important consequences from all three theoretical accounts suggest that WFTC had the consequence of increasing entire hours for working lone female parents, and the negative impact on hours on some persons predicted by the theory was outweighed by the positive impact of WFTC on other persons.

The “ on the job 16-29 hours ” arrested development efforts to demo how the inducements changed for those who were working 16-29 hours before the reform. As the panel information follows the same persons, we should be able to happen the consequence of WFTC on the persons who were working 16-29 hours before the reform. I therefore removed from the sample those who were working exterior of those hours in the pre-reform period. This resulted in a negative but non statistically important consequences in both FE theoretical accounts. This would propose that for lone female parents working between 16 and 29 hours, the income consequence was approximately the same as the permutation consequence as a consequence of WFTC or that there were loosely the same figure of people whose permutation consequence outweighed the income consequence as there were persons whose income consequence outweighed the permutation consequence. However, we are utilizing an imbalanced panel which means that the information is non needfully made up of the same people pre and station reform. If there are different people who are introduced into the panel after the WFTC reform so this will do a colored calculator. For illustration, if an unemployed lone female parent enters in the post-reform period and replaces a lone female parent who worked 16-29 hours from the pre-reform period, it will look as if WFTC has had the consequence of cut downing hours from 16-29 to zero and will bias the calculator downwards. As the working hours distribution for lone parents is skewed towards fewer to no hours it is likely that the calculator will be biased downwards. The big criterion mistake for the FE ( DD ) theoretical account could demo that the magnitude of the comparative permutation and income effects varies greatly for each person. The smaller coefficient for FE ( DDD ) implies a greater upward tendency for hours in comparing to the control group. We can utilize the FRS informations to see how the mean figure of hours for those working 16-29 both pre and station reform alterations, by utilizing a sample which merely includes those working between 16-29 hours. The Tobit theoretical account consequences suggest that the consequence of WFTC on the mean figure of hours for those working between 16-29 hours is non statistically different from nothing. These consequences are consistent with the findings of both Gregg and Harkness ( 2003 ) , utilizing leaning mark matching and Eissa and Liebman ( 1997 ) who both conclude that the effects are statistically undistinguished.

We encounter the same job of imbalanced panel informations when seeking to cipher the consequence on lone female parents working more than 30 hours. If solitary female parents enter and replace other lone female parents, unless they work the same figure of hours as the persons they replace, they will bias the consequences. Again the skewed distribution of hours worked towards fewer hours will bias the consequence downwards. The FE ( DD ) consequences imply that in response to the WFTC reform, lone female parents who were working 30 hours and over hold reduced their hours by 20 per centum on norm. These consequence is rather high, perchance due to the fact that a big proportion of lone female parents who work over 30 hours will non work really much more than 30 hours ( due to home committednesss ) , in the hours part in which the theory shows has unambiguous negative inducements to diminish work. This is besides combined with the downwards prejudice produced by the imbalanced panel. The Tobit theoretical account uses the same theoretical account as before where those working below 30 hours are removed from the sample, to give the alteration in mean hours worked for those working 30 hours and over. The consequences show a statistically important negative response where the mean figure of hours is decreased by 1.5 per centum. This figure is likely to be more accurate as it does non endure from the same prejudice as panel informations, nevertheless it can non demo if people decrease their hours below 30 hours as a consequence of WFTC. The likeliness that lone female parents decrease hours below 30 is improbable as they face the big deterrence of losing their excess child care recognition. The consequences from the employment arrested developments suggest that WFTC causes small motion up to working 30 hours or motion from over to below 30 hours. In comparing to the Tobit theoretical account consequences, the FE ( DD ) consequences appear reasonably implausible and biased.


In October 1999 WFTC replaced Family Credit to make greater inducements to work for workless or low-income families with the purpose of cut downing child poorness. I focus on the effects of the reform on individual female parents, as they were a peculiar mark, utilizing informations from the Family Resources Survey and the British Household Survey. I attempt to place the consequence of WFTC on employment and on hours worked comparing a intervention group dwelling of lone female parents with an control group dwelling of individual adult females without kids who are ineligible for WFTC. On the footing of my fixed effects arrested development utilizing the difference in difference method consequences show that WFTC increased labour-force engagement by 8.2 per centum, eligible employment by 9.6 per centum but had no statistically important consequence on full clip employment. These consequences are consistent with the theory that WFTC increases inducements to work 16 hours unequivocally, but does non significantly increase the inducements recognition to work 30 hours or over purely in comparing to FC.

My consequences besides show that WFTC increases entire hours worked by lone female parents between 15.7 per centum and 30.7 per centum depending on the theoretical account used and besides increases the mean figure of hours worked by those already in a occupation. The reform does non hold a statistically important consequence on hours for those working between 16 and 29 hours harmonizing to the consequences, nevertheless it is likely that it does bring on negative inducements to work for those working 30 hours and above, diminishing the figure of hours worked by 1.5 % harmonizing to the Tobit theoretical account consequences. These consequences are besides consistent with the theory. A more thorough probe into the effects on hours worked would necessitate a balanced panel.

The grounds suggests that WFTC has had a positive impact on employment for lone female parents while maintaining inducements to work hours reasonably changeless. This would hold redistributed income in favor of low-income lone female parents and would hold probably had a positive impact on cut downing child poorness.

  1. Department for Work and Pensions. ( 2009 ) . Households Below Average Income Statistics 2007/08.
  2. Salvage the Children. ( 2010 ) . Measuring Severe Child Poverty in the UK.
  3. Harry Wallop. ( 2009 ) .Nearly 3 million kids in poorness in Britain.Available: hypertext transfer protocol: // Last accessed 1 March 2010.
  4. OECD Economic Outlook 2001
  5. The GREGG PAPER from BB
  6. OECD, 1998, and Micklewright 2000 – demand to look this up
  7. Card and Robbins ( 1998 ) – this excessively
  8. 1988-93 Current Population Survey informations ( what is this it ‘s in C & A ; R p6 paper ) OR See Card and Robins ( 1988, Table 2 )
  9. Blundell OXREP
  10. Card and Robins Financial Incentives for Increasing Work and Income Among Low-Income Families ( 9022010 paper )
  11. Elissa and Liebman
  12. Hotz and… .
  13. Meghir revenue enhancement paper
  14. Blomquist and Newey ( 2002 )
  15. Kaiser et Al ( 1992 )
  16. Jenkins ( 1992 )
  17. Section 3

  18. Cheng Hsiao ( 2006 ) Panel Data Analysis – Advantages and Challenges – do n’t cognize if I need this.
  19. Francesconi and Van der Klauuw ( 2006 )
  20. the micro book
  1. Other income, paysheet revenue enhancements and nutrient casts are excluded from the budget line for simpleness.
  2. Exact figure can non be given as this varied immensely between US provinces.
  3. In the illustration, the chief earner worked 40 hours and the secondary earner had big deterrences to diminish hours when working under 10 hours.
  4. There is a possibility that the same figure of people started to work over 30 hours as those who started to work under 30 hours, although this is improbable and can non be explained through rational inducements.
  5. yit = & A ; szlig ; 0 + & A ; szlig ; 1Xi + & A ; szlig ; 2Tt + & A ; szlig ; 3Xi *Tt +eit
  6. The dtobit arrested development is used to happen the fringy effects given.