Unemployment Persistence, Duration Dependence, and Long-Term Unemployment: A Markov Perspective

Size: px
Start display at page:

Download "Unemployment Persistence, Duration Dependence, and Long-Term Unemployment: A Markov Perspective"

Transcription

1 Uneployent Persistence, Duration Dependence, and Long-Ter Uneployent: A Markov Perspective by George Sheldon University of Basle Labor Market and Industrial Organization Research Unit (FAI) Abstract: This paper investigates the foral link between uneployent persistence, on the one hand, and negative dependence duration and long-ter uneployent on the other. Although negative duration dependence (iplying that the long-ter uneployed exhibit a lower re-eployent probability than the short-ter jobless) frequently appears in the literature as an explanation for the persistently high levels of uneployent in Europe, a foral link has yet to be established. Modeling uneployent as an absorbing Markov chain, we deonstrate that uneployent persistence in fact does not iply negative duration dependence, nor vice versa. Uneployent persistence does however suggest a large share of long-ter uneployed. As a result, the proportion of long-ter uneployed will perfor well in uneployent tie-series regressions, whether negative duration pertains or not. Eploying data for Switzerland, the paper reveals, in addition, that the degree of uneployent persistence is not a constant, as univariate tie-series analyses custoarily assue, but rather a function of the general level of econoic activity and the duration of uneployent insurance eligibility. Acknowledgents: This work was funded by a research grant fro the Swiss Uneployent Insurance Syste, to who we wish to express our gratitude. Thanks are also due Hans Gersbach, who provided in-depth coents on an earlier draft. I have also profited fro coents fro Nicholas Kiefer, George Neuann, and Jan van Ours. The usual caveats apply, however. Keywords: persistence, duration dependence, long-ter uneployent, Markov chain, uneployent insurance JEL-Code: C41, J64, J65 1

2 1. Introduction A close relationship between uneployent persistence, duration dependence, and long-ter uneployent has established itself in the literature in the last decade. Uneployent persistence refers to the tendency of the stock of uneployed in a nuber of European countries to reain well above pre-recession levels long after a recovery fro adverse shocks has set in. As a consequence, the tie path of uneployent in the countries affected has acquired a step-like appearance, with each successive cyclical downturn pushing uneployent up another notch. This pattern is well docuented in the literature (e.g., Alogoskoufis/Manning 1988, OECD 1994). Tie series studies of uneployent (e.g. Blanchard/Suers 1986, Barro 1988) have tried to capture this pattern by odeling uneployent as a first-order autoregressive process in which the coefficient of lagged uneployent easures the degree of persistence. One reason given in the literature for uneployent persistence is the deterioration of skills through prolonged uneployent, which worsens the uneployed s chances of regaining work (e.g., Budd et al. 1987). 1 This explanation corresponds essentially to the concept of negative duration dependence (Heckan/Borjas 198), according to which a jobless person s probability of re-eployent declines with the elapsed duration of uneployent. Negative duration dependence iplies that the long-ter uneployed have a harder tie finding work than the short-ter jobless. 2 Viewed in this way, negative duration dependence sees to provide a link between persistence and long-ter uneployent. This apparent connection has led to the use of the share of long-ter uneployed as a proxy for persistence in Beveridge curve regressions, in which the level of uneployent serves as the dependent variable (e.g., Budd et al. 1987; Christl 1988; Franz 1987; Jackan et al. 1989). In the sae vein, the OECD (1983, 1987, 1994) continually points to a positive correlation, across tie and country, between the aggregate level of uneployent and the proportion of long-ter uneployed. 1 2 Other reasons appearing in the literature are: (i) reductions in the capital stock in association with adverse shocks to eployent, which reduce the subsequent deand for labor, thereby protracting uneployent (e.g., Burda 1988) and (ii) insider-outsider structures that prevent the uneployed (outsiders) fro affecting wage-bargaining outcoes supportive of their re-eployent (e.g., Blanchard/Suers 1986). Another reason for negative duration dependence, besides the deterioration of skills, is the negative signal that long-ter uneployent sends iperfectly infored firs that use the duration of uneployent as a screening device for worker productivity (e.g., Tötsch 1988). 2

3 Despite the suggestiveness of a relationship existing between uneployent persistence, on the one hand, and duration dependence and long-ter uneployent on the other, a foral link has yet to be established. This ay be due in part to the fact that these concepts pertain to different levels of aggregation: uneployent persistence relates to the tie path of aggregate uneployent, whereas duration dependence refers to the evolution of individual joblessness. The following paper offers a odeling fraework that bridges the gap between aggregate and individual uneployent and in so doing establishes a foral relationship between persistence, duration dependence, and the share of long-ter uneployed. The approach consists in odeling uneployent as an absorbing Markov chain in which duration classes serve as the transition states through which oveent occurs, and outside-of-registered-uneployent acts as the absorbing or terinal state. Using this approach, we show that contrary to received thought, uneployent persistence does not iply negative duration dependence nor vice versa. We deonstrate that persistence does however iply a large share of long-ter uneployed, which is why the level of uneployent and the share of long-ter uneployed are found to correlate positively. As a consequence, the proportion of long-ter uneployed will perfor well in tie-series regression odels of uneployent irrespective of the presence of duration dependence. We apply the Markov approach to onthly Swiss uneployent data for the period The results confir our theoretical findings pertaining to the link between persistence, duration dependence and long-ter uneployent. Moreover, our results show that in contrast to conventional tie-series analyses of uneployent, persistence is not a paraeter of the labor arket, but rather a function of the level of econoic activity and the duration of uneployent insurance eligibility. Furtherore, we find that negative duration dependence varies positively with the business cycle, being ore pronounced in upswings than in slups. The paper unfolds as follows. Section 2 offers a stylized version of the traditional ethod of odeling uneployent persistence. Section 3 copares the Markov approach with the traditional odeling procedure and describes the foral links between persistence, duration dependence and the proportion of long-ter uneployed. Section 4 presents the results fro applying the Markov odel to Swiss data. Section 5 suarizes our results and draws policy conclusions. 3

4 2. Uneployent Persistence It is coon practice in the literature (cf. Franz 199) to odel uneployent persistence essentially as a first-order non-hoogeneous linear difference equation (1) U = ρ U + x τ τ 1 τ, in which U τ represents the stock of uneployed at tie τ, x τ denotes an exogenous variable, and ρ is a constant with range [, 1]. Fro a tie-series perspective, (1) would constitute a first-order autoregressive process with the exogenous variable x τ sybolizing a white noise or oving-average process. The coefficient ρ easures the persistence of the process iplied by (1), as it deterines the speed by which U adjusts to changes in x. To deonstrate this, we regard the tie path of U between two equilibria U * (x) and U * (x+ x), where U * (x) denotes the equilibriu 3 level of the stock of uneployed associated with level x and x represents a peranent shock to x. Fro (1) it is clear that (2) ( ) and U * x = (3) ( ) U * x+ x x 1 ρ = x + x 1 ρ. We start at tie τ = and assue that a peranent shock x occurs to x. According to (1), the level of uneployent periods later will be (4) 2 1 * ( )[ 1 ρ ρ K ρ ] ρ ( ) U = x+ x U x = ( + )( 1 ρ ) x x x ρ + 1 ρ 1 ρ. The change in the level of uneployent after periods is thus equal to the difference between (2) and (4), i.e., (5) U U * ( x ) = x 1 ( ρ ) 1 ρ 3 Asterisks denote equilibriu values. 4

5 At the end of the adjustent process, uneployent will have changed in accordance with (2) and (3) in total by * * (6) U ( x x) U ( x ) + = x 1 ρ. Thus after periods, the degree α to which adjustent is coplete equals (7) α = = 1 ρ * U U ( x) ( + ) ( ) * * U x x U x. Solving for in (7) yields (8) = ln ( 1 α) ln ρ. Hence, represents the nuber of periods required for the level of uneployent to adjust to a degree α to a new value of the exogenous variable. Observe that the size of the peranent shock x has no bearing on the value of. Equation (8) clearly deonstrates that the tie required for uneployent to adjust to any given degree α to a new equilibriu level is an increasing function of ρ across the range [, 1]. In other words, large values of ρ iply that uneployent will be slow to adjust to iproved econoic conditions, while low values iply rapid convergence. If ρ =, adjustent is instantaneous, whereas it is never-ending if ρ = 1, since in this case no unique equilibriu level of uneployent exists. This latter case is often tered hysteresis in the econoic literature, while values of ρ not greatly lower than one are denoted as persistence (cf. Wyploscz 1987). An iportant concept in regard to α is the so-called edian lag, which stes fro tie-series odels (cf., e.g., Greene 1993). The edian lag gives the nuber of periods the adjustent process takes to run 5 percent of its course, i.e., it pertains to the value assues when α equals.5. In the following, we eploy the edian lag in place of ρ as our easure of uneployent persistence. Having set out the autoregressive approach, we now contrast it with the Markov odeling procedure. 5

6 3. Absorbing Markov Chain of Uneployent Basically, a Markov chain 4 is a probabilistic odel that describes the law of oveent of a variable across a finite set of utually exclusive states in discrete tie. A Markov chain is tered absorbing if at least one state exists, the absorbing state, which can ultiately be reached fro all other states but fro which there is no escape. In our application of this odeling fraework to uneployent, we equate the non-absorbing states with uneployent duration classes and the absorbing state with non-uneployent. 5 A Markov chain of uneployent subdivides the elapsed duration of ongoing spells of uneployent into a finite set of duration classes, here defined as {(, 1], (1, 2],..., (T, T+1]}, where T sybolizes the assued longest possible spell of uneployent. The definition of the set of duration classes indicates that they have unifor width equal to one unit of process tie. We also assue that entry and exit fro uneployent occurs iediately after a period ends, so that a duration class (t-1, t] encopasses all individuals in the stock of uneployed who have been jobless for t periods. 6 The central law of otion governing an absorbing Markov chain of uneployent is given by 7 (9) U = P U 1 + N, τ τ τ where U represents a colun vector reflecting the distribution of the stock of uneployed across the T+1 duration classes defined above. N is a colun vector giving the nuber of new entries into uneployent per period. Since entry into uneployent is only possible through the lowest duration class ( 1], only the first eleent in N, which we denote by N, is non-zero. As in all Markov chains, oveent through the various states (duration classes) is governed by a square atrix P of constant transition probabilities, which in our case give an uneployed person s probability of passing fro one duration class to the Snell (1988) provides a good overview of Markov chains. By contrast, previous Markov odels of labor arket transitions have, as a rule, been non-absorbing and have pertained to different states of labor arket participation (uneployent, eployent, non-participation), as opposed to different duration classes of uneployent. See, e.g., Marston (1976), Toikka (1976) and Kiefer/Neuann (1989). Note that t pertains to duration ("age"), whereas τ refers to tie. Duration and tie are easured in identical units, however. The distinction between duration and tie should becoe apparent upon viewing (22) below. Bold letters denote atrices. 6

7 next. Since every individual oves in each period (as the width of a duration class corresponds to one unit of process tie) and then either into the next higher duration class or out of uneployent altogether, only those transition probabilities lying directly above and parallel to the ain diagonal of P are non-zero. 8 Of these, a given transition probability p(t) for t = 1,..., T gives an uneployed person s probability of passing fro duration class (t-1, t] to duration class (t, t+1], i.e., of reaining uneployed a further period, conditional on having been uneployed for t periods. In other words, the transition probabilities constitute duration-class-specific survivor rates. We assue that p() = 1 and that p(t) =, i.e., that no one enters and exits uneployent within the sae period and that no spell of uneployent lasts ore than T periods. The copleent of a survivor rate is tered its hazard rate in duration analysis. 9 The sequence of hazard rates h(t) for t = 1,..., T represents the hazard function 1 of the uneployent process in discrete tie. Duration dependence refers to the slope of the hazard function. Negative (positive) duration dependence pertains if the hazard rates decrease (increase) with increasing elapsed duration t. The situation in which h(t) does not vary across duration classes is tered the exponential case, since the hazard function of an exponential distribution is equal to a constant. Note that negative duration dependence iplies positive duration dependence with respect to survivor rates. As a coparison of (9) with (1) indicates, the essential difference between the ore conventional odeling approach of Section 2 and a Markov setting is that the latter subdivides the stock of uneployed into individual duration classes and equates the exogenous variable x to inflows into uneployent. 11 Consequently, both approaches exhibit a nuber of foral siilarities. This applies, for one, to the adjustent tie path (4). In a Markov fraework, (4) is defined in accordance with (9) as U = N+ N I+ P+ P + K + P + U N P, 2 1 * (1) [ ][ ] ( ) where I sybolizes the identity atrix and U * (N), the equilibriu duration distribution associated with inflows N. Note of course that the tie path (1) pertains to the duration class distribution of the uneployed periods after a peranent shift in the size of entering cohorts fro N to N+ N, sybolizing an exogenous shock. The first Stone (1971) investigates the properties of this particular for of the transition atrix in a deographic context. Kiefer (1988) provides a good overview of the basics of duration analysis. Behaviorally, hazard functions of uneployent can be viewed as reduced fors based on job search odels. See Devine/Kiefer (1991). Treating inflows into uneployent as exogenous is also a coon procedure in deriving the Beveridge curve fro a atching function. See, e.g., Blanchard/Diaond (1989). 7

8 ter on the right-hand side of (1) refers to the reaining stock of all cohorts that entered uneployent in the periods after tie τ =, and the second ter applies to the reaining uneployed fro the stock at tie τ =. Allowing to increase to infinity in (1) yields the equilibriu or steady-state duration class distribution of the uneployed associated with inflow level (N+ N), i.e., * (11) U ( N+ N ) = [ N+ N][ I P] 1. Note the foral siilarity between (11) and (3), the latter steing fro the conventional odeling approach. The inverse atrix [I - P] -1 in (11) is tered the fundaental atrix of a Markov chain in the literature. Due to the special diagonal structure of P, the fundaental atrix of our odel consists of an upper triangular atrix in which each row, starting at the ain diagonal, represents the conditional discrete tie survivor function 12 S(t t r- 1) of those uneployed with an elapsed duration t of at least one less than the given atrix row nuber r, i.e., (12) Stt ( r 1) = t i= r 1 j= pi () pj (). The first row of the fundaental atrix thus corresponds to the conditional survivor function for entering cohorts, i.e., the unconditional survivor function S(t), where (13) St () = t pi () i=. The survivor function S(t) gives the individual probability of reaining uneployed for at least t+1 periods. It can also be regarded as the proportion of ebers of a cohort of uneployed that can expect to reain uneployed for this duration. We intend now to derive the forula for the edian lag of our Markov chain of uneployent. In order to link the Markov odel to uneployent persistence, we concentrate on the stock of uneployed, as opposed to their distribution across dura- 12 The conditional survivor function finds use in the study of job tenure, as the area under this function corresponds to the expected rest duration of an event (i.e., holding the sae job), given that it has already lasted t periods, on which job duration studies often focus (e.g., Hall 1982). 8

9 tion classes. We begin by noting that, since only the first row of the colun vector N contains a non-zero eleent, (11) reduces to * (14) U ( N+ N ) = [ N + N][ 1 p(1) p(1)p(2) K p(1)p(2) K p(t-1)p(t) ] Hence, * ( ) ( ) * U N+ N = U N+ N e (15) ( N + N )[ 1 p(1) + p(1)p(2) + K + p(1)p(2) Kp(T-1)p(T) ] = + ( ) = N+ N Si (), T i= where e denotes the unit vector. In the exponential case, (15) siplifies essentially to (3), as in this instance all survivor rates are equal. Analogous to (15), * (16) U ( N ) = N T i= S() i, so that the total change in the steady-state level of uneployent as a result of a peranent change in the size of entering cohorts of N is equal to * * (17) ( ) ( ) U N+ N U N = N S() i T i=. Furtherore, the scalar equivalent to (1) corresponds to U = N + N S () i + N S () i (18) ( ) 1 i= T i=, iplying that * (19) ( N ) U U = N S() i. 1 i= 9

10 Thus, the degree of adjustent achieved after periods is equal to the ratio of (19) to (17), so that (2) α = * U U ( N) ( + ) ( ) * * U N N U N 1 i = = T j= Si () Sj (). As (2) indicates, the edian lag (α =.5) within a Markov odel of uneployent corresponds to that value of (nuber of adjustent periods) that divides the area under the survivor function S(t) into equal halves. This result is - to our knowledge - new in the literature. Note that the edian lag does not depend on the level of inflows into uneployent nor their changes, but rather is solely a function of survivor rates. This corresponds to the finding with respect to (8) that the size of exogenous shocks has no bearing on the value of the edian lag. Nor does the edian lag depend on the gradient of the hazard function. This becoes clear by observing that (2) reduces to (7) 13 if survivor rates do not vary across duration classes (exponential case), iplying that the general level of survivor rates are also a principle factor deterining the speed of adjustent. Hence, if the survivor rates (hazard rates) are high (low) enough on average, the tie path of uneployent will exhibit persistence whether negative duration pertains or not. In other words, negative duration dependence is not a necessary condition for uneployent persistence. Nor is it a sufficient condition, since an adequately low average survivor rate will lead to a short edian lag even in the presence of negative duration dependence. This results by no eans represents a special case. As we shall see in the next section, it is not uncoon, in fact, for the degree of negative duration dependence to fall while uneployent persistence increases. A long edian lag of uneployent adjustent does however iply a large share of long-ter uneployed. To see why, observe that, based on (15) and under the assuption that a period equals one onth, the steady-state proportion of long-ter uneployed (those uneployed with elapsed durations in excess of 12 onths) is equal to With ρ replaced by the constant survival rate p, of course. Budd et al. (1988) derive (21) without reference to Markov chains. 1

11 * (21) L T i = = 12 T j= Si () Sj (). Fro (21) it is apparent that the steady-state share of long-ter uneployed corresponds to that proportion of the area below the survivor function lying above the duration class (11, 12]. Since a long edian lag iplies that a large share of the area beneath the survivor function lies in the upper tail of the function, it also suggests a large share of long-ter uneployed. The intuition behind this result is that an exogenous change in the size of inflows into uneployent will take all the longer to work its way through the stock of uneployed the larger is the share of cohort ebers that exit uneployent fro high duration classes. Since the steady-state share of long-ter uneployed is a good proxy for the edian lag of uneployent adjustent to exogenous shocks, one should expect that the ratio of long-ter uneployed to total uneployent to perfor well in tie-series regressions of uneployent, whether duration dependence pertains or not. As a proxy for the edian adjustent lag, the share of long-ter uneployed acts uch like an error correction ter, adjusting for any short-ter deviations fro a long-run relationship between uneployent and a set of exogenous variables. 4. Case Study for Switzerland In the following we exaine the relationship between uneployent persistence, duration dependence, and long-ter uneployent epirically with respect to Switzerland for the saple period The choice of Switzerland is due to data availability. The 199s brought a sharp rise to uneployent in Switzerland. Throughout the 197s and 198s the uneployent rate rarely rose above 1 percent and stood at a ere.5 at the beginning of 199 (Figure 1). Since then, however, the uneployent rate in Switzerland has increased alost tenfold rising to nearly 5 percent by the beginning of In contrast to official practice in Switzerland, the nuerator of the uneployent rate in Figure 1 equals the su of registered uneployed and total eployed, the latter drawn fro annual national eployent statistics (Erwerbstätigenstatistik) instead of fro 1-year census figures. Our ethod of calculation yields virtually identical results up to the 199s when they begin to fall short of official figures, which ignore changes in the size of the labor force between censes. Since eployent 11

12 Figure 1: Year-Beginning Uneployent Rate, Switzerland, Percent Year The tie path of uneployent in the 199s has followed rather closely the cyclical developent of deand, as easured by the relative deviation of the seasonally and shock-adjusted onthly index of industrial production fro its trend. As Figure 2 indicates, uneployent exhibits the sae double-dip pattern as deand. Cyclical downturns in uneployent are of course not new to Switzerland. In fact, in the wake of the first oil crisis in the id 197s eployent dropped by over 8 percent, the largest decline in any OECD eber country, yet uneployent failed to reach even the 1 percent ark, the reason being that in the past a large nuber of workers exited the labor arket upon being released: foreign workers with teporary work perits returned hoe while resident workers, being uninsured against job loss, chose not to register at the eployent office and thus were not counted as uneployed. Since then, however, the situation has changed. For one, copulsory uneployent insurance coverage was introduced in April 1977 and then extended in 1984, with the enactent of the national uneployent insurance law. Secondly, the share of foreign workers with a peranent residency perit and thus the right to reain in Switzerland without a job has increased greatly. As a result of these changes, a drop in uneployent today is ore fully reflected in the uneployent statistics than in the past. has risen since 199 the official rate of 5.3 percent at the start of 1997 exceeds the 4.8 percent appearing in the chart. 12

13 Figure 2: Relative Deviation of Swiss Industrial Production 16 about Its Trend, Percent Calendar Month It is interesting to note that despite the low level of uneployent that Switzerland has enjoyed in the past, the tie path of the uneployent rate nonetheless displays the step-like pattern coon to EU eber countries, suggesting that Switzerland too has suffered fro uneployent persistence. In the following epirical investigation, we apply equations (2) and (21) to Swiss uneployent data. In order to ipleent these equations, we require survivor rates for the various duration classes. We estiate onthly survivor rates for 49 duration classes with a unifor length of one onth based on the forula 1for t = (22) pt () τ = N(t) τ N(t 1) τ 1 for t = 1, K, 47 and for t = 48, where τ = January 199,..., Deceber N(t) τ represents the nuber of uneployed in the duration class (t, t+1] on the last working day of the calendar onth τ Based on the onthly industrial production index of BAK Econoics, Basle. Equation (22) is based on the so-called life-table ethod. The forula assues that no spell of registered uneployent lasts longer than 48 onths, an assuption violated by no ore than.2 percent of all spells in our saple. Furtherore, it assues that entries in and exits fro uneployent occur on the first of each onth. This assuption siplifies calculations and has been found to be true for a ajority of uneployent spells in an earlier study (Sheldon 1989). The forula also excludes individuals who enter and exit uneployent in the sae calendar onth. In so doing, (22) underestiates the true size N() of entering cohorts, but this is iaterial for the purposes of our study. 13

14 Observe that N(t) τ and N(t-1) τ-1 refer to the sae cohort, so that (22) is based on intracohort survivor rates. The values for N(t) τ are obtained fro the onthly tapes of the coputer-based placeent and eployent statistics syste (AVAM) of the Swiss Federal Office of Industry, Trade and Labor (BIGA), which since 199 records all spells of registered uneployent in Switzerland. The values for the edian lag (2) and the share of long-ter uneployed (21) calculated on the basis of (22) constitute the steady-state values that the survivor rates of the corresponding onth iply. In the case of long-ter uneployent the steadystate values represent the level the share of long-ter uneployed would assue if the labor arket conditions that held in the particular calendar onth reained constant until equilibriu is achieved. The observed share of long-ter uneployed will lag behind the steady-state value unless labor arket conditions reain unchanged long enough for uneployent to reach its stock-flow equilibriu. Figure 3 presents the results of applying (22) to the Swiss data. The three panels present the copleents to the calculated survivor rates (i.e., their hazard rates) in the for of 12-onth averages for 199, 1993 and 1996, respectively. The panels pertain to the first 24 onths of uneployent. 18 As the figures show, the gradient of the hazard function has leveled off over the saple period, and the average hazard rate has fallen. 19 In other words, the degree of negative duration dependence 2 has declined during the econoic slup in the 199s, while the average level of survivor rates has risen. The shift of the hazard rate peak fro duration class (11, 12] in 199 to duration class (18, 19] in 1993 and 1996 has to do with changes in the axiu duration of uneployent copensation, which was extended fro 5 weeks in 199 to 6 weeks in 1992 and finally to 8 weeks in 1993 (see Sheldon 1997, for details). As a statistical test of whether the gradient of the hazard function indeed decreased, we regressed the natural logarith of the integrated hazard Λ(t) for t = 1,..., 47 on a constant and the log of duration t for the three years depicted in Figure 3. The coefficient of log duration easures the degree of duration dependence. A coefficient of 1 indicates no duration dependence, while a value below (above) 1 signifies negative As equation (15) indicates, the survivor rates of the lower duration classes are the deciding factor deterining the value of the survivor function, since they doinate in the forula. The procyclical sensitivity of the slope of the hazard function to the level of deand as pictured in Figure 3 also holds for the USA (cf. Sider 1985). The duration dependence presented in Figure 3 ay in fact reflect "heterogeneity" instead of true duration dependence (cf. Borjas/Heckan 198). The distinction is iaterial for our purposes, however. 14

15 Figure 3: Hazard Rates, Switzerland, Annual Averages Per Month Rate Elapsed Une ploy ent Duration in M onths Per Month Rate Elapsed Une ploy ent Duration in M onths Per Month Rate Elapsed Une ploy ent Duration in M onths 15

16 (positive) duration dependence. The regression results (Table 1) confir that the degree of negative duration dependence decreased in 1993 and 1996 with respect to 199. Table 1: OLS Results for ln Λ(t) = α + β ln t α -1.76*** -2.6*** *** (.31) (.26) (.22) β.827***.945***.932*** (.1) (.9) (.7) R Asterisks denote statistical significance with a risk of error of less than 1% (*), 5% (**), or 1% (***). Standard errors appear in parentheses. Despite the decrease in the degree of duration dependence, the edian lag has risen and with it - as the results in Section 3 would lead one to expect - the steady-state share of long-ter uneployed. Figure 4 shows this clearly. The tie paths presented in the diagra represent centered 5-onth oving averages of our calculated figures. As can be seen, the edian lag length increased noticeably over the saple period although the degree of negative duration dependence fell, because the general level of the hazard function fell draatically. The fact that in 199 the edian lag was low despite a high degree of negative duration dependence confirs the result fro the previous section that negative duration dependence is not a sufficient condition for uneployent persistence. And the fact that after 199 the edian lag rose although duration dependence decreased clearly deonstrates that negative duration dependence is also not a necessary for uneployent persistence. Thus, Figures 3 and 4 provide epirical support for the foral results obtained in Section 3. Furtherore, the fact that the tie path of the edian lag length closely parallels that of the steady-state share of long-ter uneployed confirs, as pointed out in the previous section, that the ratio of long-ter uneployed to total uneployent reflects, in effect, the degree of uneployent persistence. This is the reason why the share of long-ter uneployed perfors well in uneployent regression equations. But since the size of the share of long-ter uneployed can, as here, be a poor indicator of the degree of negative duration dependence, uneployent regression equations that include the share of long-ter uneployed as a right-hand variable yield inconclusive results with respect to the presence of negative duration dependence Although uneployent regression equations use the observed or, fro a Markov perspective, disequilibriu share of long-ter uneployed and not the steady-state figures appearing in Figure 4, 16

17 Figure 4: Median Lag and Long-Ter Uneployent, Switzerland, Median Lag in Months Long-Ter Share Calendar Month Median Lag Long-Ter Uneployent To link the tie path of the edian lag and the steady-state share of long-ter uneployed to the level of econoic activity and to changes in the duration of uneployent insurance eligibility, we regressed the natural logariths of the non-soothed variables appearing in Figure 4 on (i) the business cycle (CYCLE) variable presented in Figure 2, (ii) a trend, (iii) onth-specific duy variables, (iv) variables capturing the age and skill coposition of the stock of uneployed, and (v) two regie duy variables UI92 and UI93 (April 1992-Deceber 1995 and April 1993-Deceber 1995, respectively) that are intended to capture the two extensions of copensation eligibility entioned above. The results pertaining to the business cycle, trend and regie variables appear in Table 2. Table 2: Median-Lag and Long-Ter Uneployent, OLS Results, Left-Hand Variable CYCLE TREND UI 92 UI 93 DW adj. R 2 ln Long-Ter Share -7.43***.9***.41***.126*** (.68) (.2) (.71) (.6) ln Median Lag -5.49***.5***.29***.17*** (.36) (.1) (.36) (.33) Asterisks denote statistical significance with a risk of error less than 1 percent (*), 5 percent (**) or 1 percent (***). Consistent standard errors appear in parentheses. this fact does not detract fro our arguent, since the left-hand variable in uneployent regressions is also a disequilibriu value, i.e., does not correspond to, say, (16). Hence, the fact that uneployent regressions are based on disequilibriu observations iplies that the results are even less conclusive than theory alone would have one believe. 17

18 According to the results in the table, the sensitivity of the edian uneployent lag and the steady-state share of long-ter uneployed to cyclical fluctuations in econoic activity is quite high. A one percentage point fall in the relative deviation of industrial production fro its trend would increase the edian lag length by 5.49 percent and the steady-state share of long-ter uneployed by 7.43 percent. According to Figure 2, it takes roughly a half a year of trend-exceeding growth of industrial production on average to achieve a one percentage point increase in the relative deviation of industrial production fro its trend level. Viewed fro this perspective, the effects of extending uneployent insurance benefit entitleents on the length of the edian lag of uneployent adjustent and on the steady-state share of long-ter uneployed are quite substantial. Based on Table 2, the extension in 1992 raised the edian lag by 23 percent and the share of long-ter uneployed by 51 percent, while the additional extension in 1993 increased both variables by 11 percent and 13 percent, respectively. 22 The total UI-induced increase in the edian lag length and in the share of long-ter uneployed corresponds in absolute size to the effect of 1½-2 years, respectively, of trend-exceeding growth in anufacturing. In other words, the entitleent extensions had a quite substantial negative ipact on uneployent persistence. 5. Suary and Conclusions The paper has shown that, contrary to popular thinking, uneployent persistence, understood as a slow adjustent speed (edian lag) of the stock of uneployed to exogenous shocks, does not iply negative duration dependence, nor vice versa. Rather, uneployent persistence depends principally on the general level of survivor rates, not on their gradient across duration classes. Hence, negative duration dependence does not explain persistently high levels of uneployent in Europe. In fact in Switzerland, negative duration dependence declines during econoic downturns. This cyclical variability of the slope of the hazard curve iplies that proportional hazard odels, which have coe to doinate the econoetric duration analysis literature and which assue unifor proportional effects across all hazard rates, are is-specified with respect to tie-varying business-cycle variables. On the other hand, uneployent persistence does iply a large share of long-ter uneployed, the reason being that both concepts depend on the relative distribution of 22 Note that the coefficients of duy variables in a logarithic regression equation easure first differences of logs. Hence, exp(β)-1, not β, gives the percentage change of the left-hand variable in response to a regie shift. 18

19 the ass of the survivor function. If a large share of ass is concentrated in the upper tail of the survivor function, the edian adjustent lag of uneployent to exogenous shocks will be long and the steady-state share of long-ter uneployed large. Hence, the share of long-ter uneployed, acting uch like an error-correction ter, will perfor well in Beveridge regressions whether negative duration dependence holds or not. Furtherore, uneployent persistence is not a constant as autoregressive odels of uneployent iplicitly assue. Rather in the case of Switzerland, persistence varies with deand conditions and the length of uneployent benefit entitleents. In fact, the negative ipact of entitleent extensions on uneployent persistence was found to be quite substantial, corresponding in absolute ters to the effect of 1½ years of trend-exceeding growth in anufacturing. The policy iplications of our results appear rather clear-cut. In order to avoid uneployent persistence, policy easures ust be adopted that avoid long spells of uneployent. Based on our findings, long-ter entitleents to uneployent insurance benefits are counterproductive. They increase the share of long-ter uneployed, thus slowing the response of uneployent to econoic recovery. 19

20 References Alogoskoufis, G., Manning, A. (1988), "Uneployent Persistence," in: Econoic Policy, pp Barro, R., (1988), "The Persistence of Uneployent," in: Aerican Econoic Review, Papers and Proceedings, Vol. 78, pp Blanchard, O., Suers, L. (1986), "Hysteresis and the European Uneployent Proble," in: S. Fisher (ed.), Macroeconoics Annual 1986, NBER, Cabridge (Mass.), pp Blanchard, O., Diaond, P. (1989), "The Beveridge Curve," in: Brookings Papers on Econoic Activity, No.1, pp Budd, A., Levine, P., Sith, P. (1987), "Long-Ter Uneployent and the Shifting U-V Curve. A Multi-Country Study," in: European Econoic Review, Vol. 31, pp Budd, A., Levine, P., Sith, P. (1988), "Uneployent, Vacancies and the Long- Ter Uneployed," in: Econoic Journal, Vol. 98, pp Burda, M. (1988), "Is There a Capital Shortage in Europe," in: Weltwirtschaftliches Archiv, Vol. 124, pp Christl, J. (1988), "An Epirical Analysis of the Austrian Beveridge Curve," Epirica - Austrian Econoic Papers, pp Devine, T., Kiefer, N. (1991), Epirical Labor Econoics. The Search Approach, OUP, New York. Franz, W. (1987), "Hysteresis, Persistence and the NAIRU: An Epirical Analysis for the Federal Republic of Gerany," in: R. Layard, L. Calfors (eds.), The Fight against Uneployent, MIT Press, Cabridge (Mass.), pp Franz, W. (199), "Hysteresis in Econoic Relationships: An Overview," in: W. Franz (ed.), Hysteresis Effects in Econoic Models, Physica-Verlag, Heidelberg, pp Greene, W. (1993), Econoetric Analysis, 2nd edition, Macillan, New York. Hall, R. (1982), "The Iportance of Lifetie Jobs in the U.S. Econoy," in: Aerican Econoic Review, Vol. 84, pp Heckan, J., Borjas, G. (198),"Does Uneployent Cause Future Uneployent? Definitions, Questions and Answers fro a Continuous Tie Model of Heterogeneity and State Dependence," in: Econoica, Vol. 47, pp Jackan, R., Layard, R., Pissarides, C. (1989), "On Vacancies," Oxford Bulletin of Econoics and Statistics, Vol. 51., pp Kiefer, N. (1988), "Econoic Duration Data and Hazard Functions," in: Journal of Econoic Literature, Vol. 26, pp Kiefer, N., Neuann, G. (1989), Search Models and Applied Labor Econoics, CUP, Cabridge (UK). Marston, St. (1976), "Eployent Instability and High Uneployent Rates," in: Brookings Papers on Econoic Activity, , pp

21 OECD (1983), Eployent Outlook, Paris. OECD (1987), Eployent Outlook, Paris. OECD (1994), Jobs Study, Paris. Sheldon, G. (1989), Dynaik der Arbeitslosigkeit in der Schweiz, Paul Haupt Verlag, Berne/Stuttgart. Sheldon, G. (1997), "Uneployent and Uneployent Insurance in Switzerland," in: Ph. Bacchetta, W. Wasserfallen (eds.), Econoic Policy in Switzerland, Macillan, London, pp Sider, H. (1985), "Uneployent Duration and Incidence: ," Aerican Econoic Review, Vol. 75, pp Snell, L. (1988), Introduction to Probability, New York. Stone, R. (1971), Deographic Accounting and Model-Building, OECD, Paris. Toikka, (1976),"A Markovian Model of Labor Market Decisions by Workers," Aerican Econoic Review, Vol. 66, pp Tötsch, E. (1988), "Screening in Labour Markets with Heterogeneous Workers," in: R. Cross (ed.), Uneployent, Hysteresis and the Natural Rate Hypothesis, Basil Blackwell, Oxford, pp Wyploscz, C. (1987), "Coent on W. Franz: 'Hysteresis, Persistence and the NAIRU: An Epirical Analysis for the Federal Republic of Gerany'," in: R. Layard, L. Calfors (eds.), The Fight against Uneployent, MIT Press, Cabridge (Mass.), pp

OBJECTIVES INTRODUCTION

OBJECTIVES INTRODUCTION M7 Chapter 3 Section 1 OBJECTIVES Suarize data using easures of central tendency, such as the ean, edian, ode, and idrange. Describe data using the easures of variation, such as the range, variance, and

More information

Chapter 6: Economic Inequality

Chapter 6: Economic Inequality Chapter 6: Econoic Inequality We are interested in inequality ainly for two reasons: First, there are philosophical and ethical grounds for aversion to inequality per se. Second, even if we are not interested

More information

Modeling the Structural Shifts in Real Exchange Rate with Cubic Spline Regression (CSR). Turkey

Modeling the Structural Shifts in Real Exchange Rate with Cubic Spline Regression (CSR). Turkey International Journal of Business and Social Science Vol. 2 No. 17 www.ijbssnet.co Modeling the Structural Shifts in Real Exchange Rate with Cubic Spline Regression (CSR). Turkey 1987-2008 Dr. Bahar BERBEROĞLU

More information

Deflation of the I-O Series Some Technical Aspects. Giorgio Rampa University of Genoa April 2007

Deflation of the I-O Series Some Technical Aspects. Giorgio Rampa University of Genoa April 2007 Deflation of the I-O Series 1959-2. Soe Technical Aspects Giorgio Rapa University of Genoa g.rapa@unige.it April 27 1. Introduction The nuber of sectors is 42 for the period 1965-2 and 38 for the initial

More information

Estimation of Korean Monthly GDP with Mixed-Frequency Data using an Unobserved Component Error Correction Model

Estimation of Korean Monthly GDP with Mixed-Frequency Data using an Unobserved Component Error Correction Model 100Econoic Papers Vol.11 No.1 Estiation of Korean Monthly GDP with Mixed-Frequency Data using an Unobserved Coponent Error Correction Model Ki-Ho Ki* Abstract Since GDP is announced on a quarterly basis,

More information

An Extension to the Tactical Planning Model for a Job Shop: Continuous-Time Control

An Extension to the Tactical Planning Model for a Job Shop: Continuous-Time Control An Extension to the Tactical Planning Model for a Job Shop: Continuous-Tie Control Chee Chong. Teo, Rohit Bhatnagar, and Stephen C. Graves Singapore-MIT Alliance, Nanyang Technological Univ., and Massachusetts

More information

Monetary Policy Effectiveness in a Dynamic AS/AD Model with Sticky Wages

Monetary Policy Effectiveness in a Dynamic AS/AD Model with Sticky Wages Monetary Policy Effectiveness in a Dynaic AS/AD Model with Sticky Wages H J Departent of Econoics University of Copenhagen Version 1.0, April 12, 2012 Teaching note for M.Sc. course on "Monetary Econoics:

More information

COS 424: Interacting with Data. Written Exercises

COS 424: Interacting with Data. Written Exercises COS 424: Interacting with Data Hoework #4 Spring 2007 Regression Due: Wednesday, April 18 Written Exercises See the course website for iportant inforation about collaboration and late policies, as well

More information

Block designs and statistics

Block designs and statistics Bloc designs and statistics Notes for Math 447 May 3, 2011 The ain paraeters of a bloc design are nuber of varieties v, bloc size, nuber of blocs b. A design is built on a set of v eleents. Each eleent

More information

The proofs of Theorem 1-3 are along the lines of Wied and Galeano (2013).

The proofs of Theorem 1-3 are along the lines of Wied and Galeano (2013). A Appendix: Proofs The proofs of Theore 1-3 are along the lines of Wied and Galeano (2013) Proof of Theore 1 Let D[d 1, d 2 ] be the space of càdlàg functions on the interval [d 1, d 2 ] equipped with

More information

Sharp Time Data Tradeoffs for Linear Inverse Problems

Sharp Time Data Tradeoffs for Linear Inverse Problems Sharp Tie Data Tradeoffs for Linear Inverse Probles Saet Oyak Benjain Recht Mahdi Soltanolkotabi January 016 Abstract In this paper we characterize sharp tie-data tradeoffs for optiization probles used

More information

Soft Computing Techniques Help Assign Weights to Different Factors in Vulnerability Analysis

Soft Computing Techniques Help Assign Weights to Different Factors in Vulnerability Analysis Soft Coputing Techniques Help Assign Weights to Different Factors in Vulnerability Analysis Beverly Rivera 1,2, Irbis Gallegos 1, and Vladik Kreinovich 2 1 Regional Cyber and Energy Security Center RCES

More information

The S-curve Behaviour of the Trade Balance: A Stepwise Procedure

The S-curve Behaviour of the Trade Balance: A Stepwise Procedure Article The S-curve Behaviour of the Trade Balance: A Stepwise Procedure Foreign Trade Review 52(1) 1 14 2017 Indian Institute of Foreign Trade SAGE Publications sagepub.in/hoe.nav DOI: 10.1177/0015732516650826

More information

About the definition of parameters and regimes of active two-port networks with variable loads on the basis of projective geometry

About the definition of parameters and regimes of active two-port networks with variable loads on the basis of projective geometry About the definition of paraeters and regies of active two-port networks with variable loads on the basis of projective geoetry PENN ALEXANDR nstitute of Electronic Engineering and Nanotechnologies "D

More information

The Transactional Nature of Quantum Information

The Transactional Nature of Quantum Information The Transactional Nature of Quantu Inforation Subhash Kak Departent of Coputer Science Oklahoa State University Stillwater, OK 7478 ABSTRACT Inforation, in its counications sense, is a transactional property.

More information

THE EFFECT OF SOLID PARTICLE SIZE UPON TIME AND SEDIMENTATION RATE

THE EFFECT OF SOLID PARTICLE SIZE UPON TIME AND SEDIMENTATION RATE Bulletin of the Transilvania University of Braşov Series II: Forestry Wood Industry Agricultural Food Engineering Vol. 5 (54) No. 1-1 THE EFFECT OF SOLID PARTICLE SIZE UPON TIME AND SEDIMENTATION RATE

More information

An Approximate Model for the Theoretical Prediction of the Velocity Increase in the Intermediate Ballistics Period

An Approximate Model for the Theoretical Prediction of the Velocity Increase in the Intermediate Ballistics Period An Approxiate Model for the Theoretical Prediction of the Velocity... 77 Central European Journal of Energetic Materials, 205, 2(), 77-88 ISSN 2353-843 An Approxiate Model for the Theoretical Prediction

More information

Monetary Policy E ectiveness in a Dynamic AS/AD Model with Sticky Wages

Monetary Policy E ectiveness in a Dynamic AS/AD Model with Sticky Wages Monetary Policy E ectiveness in a Dynaic AS/AD Model with Sticky Wages Henrik Jensen Departent of Econoics University of Copenhagen y Version.0, April 2, 202 Teaching note for M.Sc. course on "Monetary

More information

Revealed Preference with Stochastic Demand Correspondence

Revealed Preference with Stochastic Demand Correspondence Revealed Preference with Stochastic Deand Correspondence Indraneel Dasgupta School of Econoics, University of Nottingha, Nottingha NG7 2RD, UK. E-ail: indraneel.dasgupta@nottingha.ac.uk Prasanta K. Pattanaik

More information

Optimal Pigouvian Taxation when Externalities Affect Demand

Optimal Pigouvian Taxation when Externalities Affect Demand Optial Pigouvian Taxation when Externalities Affect Deand Enda Patrick Hargaden Departent of Econoics University of Michigan enda@uich.edu Version of August 2, 2015 Abstract Purchasing a network good such

More information

W-BASED VS LATENT VARIABLES SPATIAL AUTOREGRESSIVE MODELS: EVIDENCE FROM MONTE CARLO SIMULATIONS

W-BASED VS LATENT VARIABLES SPATIAL AUTOREGRESSIVE MODELS: EVIDENCE FROM MONTE CARLO SIMULATIONS W-BASED VS LATENT VARIABLES SPATIAL AUTOREGRESSIVE MODELS: EVIDENCE FROM MONTE CARLO SIMULATIONS. Introduction When it coes to applying econoetric odels to analyze georeferenced data, researchers are well

More information

What is Probability? (again)

What is Probability? (again) INRODUCTION TO ROBBILITY Basic Concepts and Definitions n experient is any process that generates well-defined outcoes. Experient: Record an age Experient: Toss a die Experient: Record an opinion yes,

More information

Now multiply the left-hand-side by ω and the right-hand side by dδ/dt (recall ω= dδ/dt) to get:

Now multiply the left-hand-side by ω and the right-hand side by dδ/dt (recall ω= dδ/dt) to get: Equal Area Criterion.0 Developent of equal area criterion As in previous notes, all powers are in per-unit. I want to show you the equal area criterion a little differently than the book does it. Let s

More information

A Simplified Analytical Approach for Efficiency Evaluation of the Weaving Machines with Automatic Filling Repair

A Simplified Analytical Approach for Efficiency Evaluation of the Weaving Machines with Automatic Filling Repair Proceedings of the 6th SEAS International Conference on Siulation, Modelling and Optiization, Lisbon, Portugal, Septeber -4, 006 0 A Siplified Analytical Approach for Efficiency Evaluation of the eaving

More information

On the Macroeconomic and Welfare Effects of Illegal Immigration

On the Macroeconomic and Welfare Effects of Illegal Immigration MPRA Munich Personal RePEc Archive On the Macroeconoic and Welfare Effects of Illegal Iigration Xiangbo Liu 28. May 2009 Online at http://pra.ub.uni-uenchen.de/5469/ MPRA Paper No. 5469, posted. June 2009

More information

Multi-Dimensional Hegselmann-Krause Dynamics

Multi-Dimensional Hegselmann-Krause Dynamics Multi-Diensional Hegselann-Krause Dynaics A. Nedić Industrial and Enterprise Systes Engineering Dept. University of Illinois Urbana, IL 680 angelia@illinois.edu B. Touri Coordinated Science Laboratory

More information

Model Fitting. CURM Background Material, Fall 2014 Dr. Doreen De Leon

Model Fitting. CURM Background Material, Fall 2014 Dr. Doreen De Leon Model Fitting CURM Background Material, Fall 014 Dr. Doreen De Leon 1 Introduction Given a set of data points, we often want to fit a selected odel or type to the data (e.g., we suspect an exponential

More information

ANALYSIS OF HALL-EFFECT THRUSTERS AND ION ENGINES FOR EARTH-TO-MOON TRANSFER

ANALYSIS OF HALL-EFFECT THRUSTERS AND ION ENGINES FOR EARTH-TO-MOON TRANSFER IEPC 003-0034 ANALYSIS OF HALL-EFFECT THRUSTERS AND ION ENGINES FOR EARTH-TO-MOON TRANSFER A. Bober, M. Guelan Asher Space Research Institute, Technion-Israel Institute of Technology, 3000 Haifa, Israel

More information

MSEC MODELING OF DEGRADATION PROCESSES TO OBTAIN AN OPTIMAL SOLUTION FOR MAINTENANCE AND PERFORMANCE

MSEC MODELING OF DEGRADATION PROCESSES TO OBTAIN AN OPTIMAL SOLUTION FOR MAINTENANCE AND PERFORMANCE Proceeding of the ASME 9 International Manufacturing Science and Engineering Conference MSEC9 October 4-7, 9, West Lafayette, Indiana, USA MSEC9-8466 MODELING OF DEGRADATION PROCESSES TO OBTAIN AN OPTIMAL

More information

Ensemble Based on Data Envelopment Analysis

Ensemble Based on Data Envelopment Analysis Enseble Based on Data Envelopent Analysis So Young Sohn & Hong Choi Departent of Coputer Science & Industrial Systes Engineering, Yonsei University, Seoul, Korea Tel) 82-2-223-404, Fax) 82-2- 364-7807

More information

LONG-TERM PREDICTIVE VALUE INTERVAL WITH THE FUZZY TIME SERIES

LONG-TERM PREDICTIVE VALUE INTERVAL WITH THE FUZZY TIME SERIES Journal of Marine Science and Technology, Vol 19, No 5, pp 509-513 (2011) 509 LONG-TERM PREDICTIVE VALUE INTERVAL WITH THE FUZZY TIME SERIES Ming-Tao Chou* Key words: fuzzy tie series, fuzzy forecasting,

More information

A note on the multiplication of sparse matrices

A note on the multiplication of sparse matrices Cent. Eur. J. Cop. Sci. 41) 2014 1-11 DOI: 10.2478/s13537-014-0201-x Central European Journal of Coputer Science A note on the ultiplication of sparse atrices Research Article Keivan Borna 12, Sohrab Aboozarkhani

More information

Ufuk Demirci* and Feza Kerestecioglu**

Ufuk Demirci* and Feza Kerestecioglu** 1 INDIRECT ADAPTIVE CONTROL OF MISSILES Ufuk Deirci* and Feza Kerestecioglu** *Turkish Navy Guided Missile Test Station, Beykoz, Istanbul, TURKEY **Departent of Electrical and Electronics Engineering,

More information

E0 370 Statistical Learning Theory Lecture 6 (Aug 30, 2011) Margin Analysis

E0 370 Statistical Learning Theory Lecture 6 (Aug 30, 2011) Margin Analysis E0 370 tatistical Learning Theory Lecture 6 (Aug 30, 20) Margin Analysis Lecturer: hivani Agarwal cribe: Narasihan R Introduction In the last few lectures we have seen how to obtain high confidence bounds

More information

Formal Education Versus Learning-by-Doing

Formal Education Versus Learning-by-Doing DISCUSSION PAPER SERIES IZA DP No. 8341 Foral Education Versus Learning-by-Doing Frédéric Gavrel Isabelle Lebon Thérèse Rebière July 2014 Forschungsinstitut zur Zukunft der Arbeit Institute for the Study

More information

A Simple Regression Problem

A Simple Regression Problem A Siple Regression Proble R. M. Castro March 23, 2 In this brief note a siple regression proble will be introduced, illustrating clearly the bias-variance tradeoff. Let Y i f(x i ) + W i, i,..., n, where

More information

2 Q 10. Likewise, in case of multiple particles, the corresponding density in 2 must be averaged over all

2 Q 10. Likewise, in case of multiple particles, the corresponding density in 2 must be averaged over all Lecture 6 Introduction to kinetic theory of plasa waves Introduction to kinetic theory So far we have been odeling plasa dynaics using fluid equations. The assuption has been that the pressure can be either

More information

DEPARTMENT OF ECONOMETRICS AND BUSINESS STATISTICS

DEPARTMENT OF ECONOMETRICS AND BUSINESS STATISTICS ISSN 1440-771X AUSTRALIA DEPARTMENT OF ECONOMETRICS AND BUSINESS STATISTICS An Iproved Method for Bandwidth Selection When Estiating ROC Curves Peter G Hall and Rob J Hyndan Working Paper 11/00 An iproved

More information

A method to determine relative stroke detection efficiencies from multiplicity distributions

A method to determine relative stroke detection efficiencies from multiplicity distributions A ethod to deterine relative stroke detection eiciencies ro ultiplicity distributions Schulz W. and Cuins K. 2. Austrian Lightning Detection and Inoration Syste (ALDIS), Kahlenberger Str.2A, 90 Vienna,

More information

Hierarchical central place system and agglomeration economies on households

Hierarchical central place system and agglomeration economies on households Hierarchical central place syste and aggloeration econoies on households Daisuke Nakaura, Departent of International Liberal Arts, Fukuoka Woen s University Executive suary Central place theory shows that

More information

8.1 Force Laws Hooke s Law

8.1 Force Laws Hooke s Law 8.1 Force Laws There are forces that don't change appreciably fro one instant to another, which we refer to as constant in tie, and forces that don't change appreciably fro one point to another, which

More information

Forecast Evaluation: A Likelihood Scoring Method. by Matthew A. Diersen and Mark R. Manfredo

Forecast Evaluation: A Likelihood Scoring Method. by Matthew A. Diersen and Mark R. Manfredo Forecast Evaluation: A Likelihood Scoring Method by Matthew A. Diersen and Mark R. Manfredo Suggested citation forat: Diersen, M. A., and M. R. Manfredo. 1998. Forecast Evaluation: A Likelihood Scoring

More information

13.2 Fully Polynomial Randomized Approximation Scheme for Permanent of Random 0-1 Matrices

13.2 Fully Polynomial Randomized Approximation Scheme for Permanent of Random 0-1 Matrices CS71 Randoness & Coputation Spring 018 Instructor: Alistair Sinclair Lecture 13: February 7 Disclaier: These notes have not been subjected to the usual scrutiny accorded to foral publications. They ay

More information

Keywords: Estimator, Bias, Mean-squared error, normality, generalized Pareto distribution

Keywords: Estimator, Bias, Mean-squared error, normality, generalized Pareto distribution Testing approxiate norality of an estiator using the estiated MSE and bias with an application to the shape paraeter of the generalized Pareto distribution J. Martin van Zyl Abstract In this work the norality

More information

A proposal for a First-Citation-Speed-Index Link Peer-reviewed author version

A proposal for a First-Citation-Speed-Index Link Peer-reviewed author version A proposal for a First-Citation-Speed-Index Link Peer-reviewed author version Made available by Hasselt University Library in Docuent Server@UHasselt Reference (Published version): EGGHE, Leo; Bornann,

More information

Analyzing Simulation Results

Analyzing Simulation Results Analyzing Siulation Results Dr. John Mellor-Cruey Departent of Coputer Science Rice University johnc@cs.rice.edu COMP 528 Lecture 20 31 March 2005 Topics for Today Model verification Model validation Transient

More information

Kinetic Theory of Gases: Elementary Ideas

Kinetic Theory of Gases: Elementary Ideas Kinetic Theory of Gases: Eleentary Ideas 17th February 2010 1 Kinetic Theory: A Discussion Based on a Siplified iew of the Motion of Gases 1.1 Pressure: Consul Engel and Reid Ch. 33.1) for a discussion

More information

Sexually Transmitted Diseases VMED 5180 September 27, 2016

Sexually Transmitted Diseases VMED 5180 September 27, 2016 Sexually Transitted Diseases VMED 518 Septeber 27, 216 Introduction Two sexually-transitted disease (STD) odels are presented below. The irst is a susceptibleinectious-susceptible (SIS) odel (Figure 1)

More information

Feature Extraction Techniques

Feature Extraction Techniques Feature Extraction Techniques Unsupervised Learning II Feature Extraction Unsupervised ethods can also be used to find features which can be useful for categorization. There are unsupervised ethods that

More information

Online Publication Date: 19 April 2012 Publisher: Asian Economic and Social Society

Online Publication Date: 19 April 2012 Publisher: Asian Economic and Social Society Online Publication Date: April Publisher: Asian Econoic and Social Society Using Entropy Working Correlation Matrix in Generalized Estiating Equation for Stock Price Change Model Serpilılıç (Faculty of

More information

Proc. of the IEEE/OES Seventh Working Conference on Current Measurement Technology UNCERTAINTIES IN SEASONDE CURRENT VELOCITIES

Proc. of the IEEE/OES Seventh Working Conference on Current Measurement Technology UNCERTAINTIES IN SEASONDE CURRENT VELOCITIES Proc. of the IEEE/OES Seventh Working Conference on Current Measureent Technology UNCERTAINTIES IN SEASONDE CURRENT VELOCITIES Belinda Lipa Codar Ocean Sensors 15 La Sandra Way, Portola Valley, CA 98 blipa@pogo.co

More information

Ocean 420 Physical Processes in the Ocean Project 1: Hydrostatic Balance, Advection and Diffusion Answers

Ocean 420 Physical Processes in the Ocean Project 1: Hydrostatic Balance, Advection and Diffusion Answers Ocean 40 Physical Processes in the Ocean Project 1: Hydrostatic Balance, Advection and Diffusion Answers 1. Hydrostatic Balance a) Set all of the levels on one of the coluns to the lowest possible density.

More information

DESIGN OF THE DIE PROFILE FOR THE INCREMENTAL RADIAL FORGING PROCESS *

DESIGN OF THE DIE PROFILE FOR THE INCREMENTAL RADIAL FORGING PROCESS * IJST, Transactions of Mechanical Engineering, Vol. 39, No. M1, pp 89-100 Printed in The Islaic Republic of Iran, 2015 Shira University DESIGN OF THE DIE PROFILE FOR THE INCREMENTAL RADIAL FORGING PROCESS

More information

Solutions of some selected problems of Homework 4

Solutions of some selected problems of Homework 4 Solutions of soe selected probles of Hoework 4 Sangchul Lee May 7, 2018 Proble 1 Let there be light A professor has two light bulbs in his garage. When both are burned out, they are replaced, and the next

More information

Kinetic Theory of Gases: Elementary Ideas

Kinetic Theory of Gases: Elementary Ideas Kinetic Theory of Gases: Eleentary Ideas 9th February 011 1 Kinetic Theory: A Discussion Based on a Siplified iew of the Motion of Gases 1.1 Pressure: Consul Engel and Reid Ch. 33.1) for a discussion of

More information

Interactive Markov Models of Evolutionary Algorithms

Interactive Markov Models of Evolutionary Algorithms Cleveland State University EngagedScholarship@CSU Electrical Engineering & Coputer Science Faculty Publications Electrical Engineering & Coputer Science Departent 2015 Interactive Markov Models of Evolutionary

More information

ASSIGNMENT BOOKLET Bachelor s Degree Programme (B.Sc./B.A./B.Com.) MATHEMATICAL MODELLING

ASSIGNMENT BOOKLET Bachelor s Degree Programme (B.Sc./B.A./B.Com.) MATHEMATICAL MODELLING ASSIGNMENT BOOKLET Bachelor s Degree Prograe (B.Sc./B.A./B.Co.) MTE-14 MATHEMATICAL MODELLING Valid fro 1 st January, 18 to 1 st Deceber, 18 It is copulsory to subit the Assignent before filling in the

More information

Inflation Forecasts: An Empirical Re-examination. Swarna B. Dutt University of West Georgia. Dipak Ghosh Emporia State University

Inflation Forecasts: An Empirical Re-examination. Swarna B. Dutt University of West Georgia. Dipak Ghosh Emporia State University Southwest Business and Econoics Journal/2006-2007 Inflation Forecasts: An Epirical Re-exaination Swarna B. Dutt University of West Georgia Dipak Ghosh Eporia State University Abstract Inflation forecasts

More information

This model assumes that the probability of a gap has size i is proportional to 1/i. i.e., i log m e. j=1. E[gap size] = i P r(i) = N f t.

This model assumes that the probability of a gap has size i is proportional to 1/i. i.e., i log m e. j=1. E[gap size] = i P r(i) = N f t. CS 493: Algoriths for Massive Data Sets Feb 2, 2002 Local Models, Bloo Filter Scribe: Qin Lv Local Models In global odels, every inverted file entry is copressed with the sae odel. This work wells when

More information

Question 1. [14 Marks]

Question 1. [14 Marks] 6 Question 1. [14 Marks] R r T! A string is attached to the dru (radius r) of a spool (radius R) as shown in side and end views here. (A spool is device for storing string, thread etc.) A tension T is

More information

Intelligent Systems: Reasoning and Recognition. Perceptrons and Support Vector Machines

Intelligent Systems: Reasoning and Recognition. Perceptrons and Support Vector Machines Intelligent Systes: Reasoning and Recognition Jaes L. Crowley osig 1 Winter Seester 2018 Lesson 6 27 February 2018 Outline Perceptrons and Support Vector achines Notation...2 Linear odels...3 Lines, Planes

More information

Experimental Design For Model Discrimination And Precise Parameter Estimation In WDS Analysis

Experimental Design For Model Discrimination And Precise Parameter Estimation In WDS Analysis City University of New York (CUNY) CUNY Acadeic Works International Conference on Hydroinforatics 8-1-2014 Experiental Design For Model Discriination And Precise Paraeter Estiation In WDS Analysis Giovanna

More information

Reading from Young & Freedman: For this topic, read the introduction to chapter 25 and sections 25.1 to 25.3 & 25.6.

Reading from Young & Freedman: For this topic, read the introduction to chapter 25 and sections 25.1 to 25.3 & 25.6. PHY10 Electricity Topic 6 (Lectures 9 & 10) Electric Current and Resistance n this topic, we will cover: 1) Current in a conductor ) Resistivity 3) Resistance 4) Oh s Law 5) The Drude Model of conduction

More information

National 5 Summary Notes

National 5 Summary Notes North Berwick High School Departent of Physics National 5 Suary Notes Unit 3 Energy National 5 Physics: Electricity and Energy 1 Throughout the Course, appropriate attention should be given to units, prefixes

More information

Measuring Temperature with a Silicon Diode

Measuring Temperature with a Silicon Diode Measuring Teperature with a Silicon Diode Due to the high sensitivity, nearly linear response, and easy availability, we will use a 1N4148 diode for the teperature transducer in our easureents 10 Analysis

More information

Revealed Preference and Stochastic Demand Correspondence: A Unified Theory

Revealed Preference and Stochastic Demand Correspondence: A Unified Theory Revealed Preference and Stochastic Deand Correspondence: A Unified Theory Indraneel Dasgupta School of Econoics, University of Nottingha, Nottingha NG7 2RD, UK. E-ail: indraneel.dasgupta@nottingha.ac.uk

More information

26 Impulse and Momentum

26 Impulse and Momentum 6 Ipulse and Moentu First, a Few More Words on Work and Energy, for Coparison Purposes Iagine a gigantic air hockey table with a whole bunch of pucks of various asses, none of which experiences any friction

More information

International Scientific and Technological Conference EXTREME ROBOTICS October 8-9, 2015, Saint-Petersburg, Russia

International Scientific and Technological Conference EXTREME ROBOTICS October 8-9, 2015, Saint-Petersburg, Russia International Scientific and Technological Conference EXTREME ROBOTICS October 8-9, 215, Saint-Petersburg, Russia LEARNING MOBILE ROBOT BASED ON ADAPTIVE CONTROLLED MARKOV CHAINS V.Ya. Vilisov University

More information

Non-Parametric Non-Line-of-Sight Identification 1

Non-Parametric Non-Line-of-Sight Identification 1 Non-Paraetric Non-Line-of-Sight Identification Sinan Gezici, Hisashi Kobayashi and H. Vincent Poor Departent of Electrical Engineering School of Engineering and Applied Science Princeton University, Princeton,

More information

paper prepared for the 1996 PTRC Conference, September 2-6, Brunel University, UK ON THE CALIBRATION OF THE GRAVITY MODEL

paper prepared for the 1996 PTRC Conference, September 2-6, Brunel University, UK ON THE CALIBRATION OF THE GRAVITY MODEL paper prepared for the 1996 PTRC Conference, Septeber 2-6, Brunel University, UK ON THE CALIBRATION OF THE GRAVITY MODEL Nanne J. van der Zijpp 1 Transportation and Traffic Engineering Section Delft University

More information

MASSACHUSETTS INSTITUTE OF TECHNOLOGY Department of Mechanical Engineering 2.010: Systems Modeling and Dynamics III. Final Examination Review Problems

MASSACHUSETTS INSTITUTE OF TECHNOLOGY Department of Mechanical Engineering 2.010: Systems Modeling and Dynamics III. Final Examination Review Problems ASSACHUSETTS INSTITUTE OF TECHNOLOGY Departent of echanical Engineering 2.010: Systes odeling and Dynaics III Final Eaination Review Probles Fall 2000 Good Luck And have a great winter break! page 1 Proble

More information

Mathematical Models to Determine Stable Behavior of Complex Systems

Mathematical Models to Determine Stable Behavior of Complex Systems Journal of Physics: Conference Series PAPER OPEN ACCESS Matheatical Models to Deterine Stable Behavior of Coplex Systes To cite this article: V I Suin et al 08 J. Phys.: Conf. Ser. 05 0336 View the article

More information

Measures of average are called measures of central tendency and include the mean, median, mode, and midrange.

Measures of average are called measures of central tendency and include the mean, median, mode, and midrange. CHAPTER 3 Data Description Objectives Suarize data using easures of central tendency, such as the ean, edian, ode, and idrange. Describe data using the easures of variation, such as the range, variance,

More information

CSE525: Randomized Algorithms and Probabilistic Analysis May 16, Lecture 13

CSE525: Randomized Algorithms and Probabilistic Analysis May 16, Lecture 13 CSE55: Randoied Algoriths and obabilistic Analysis May 6, Lecture Lecturer: Anna Karlin Scribe: Noah Siegel, Jonathan Shi Rando walks and Markov chains This lecture discusses Markov chains, which capture

More information

CHAPTER 19: Single-Loop IMC Control

CHAPTER 19: Single-Loop IMC Control When I coplete this chapter, I want to be able to do the following. Recognize that other feedback algoriths are possible Understand the IMC structure and how it provides the essential control features

More information

Data-Driven Imaging in Anisotropic Media

Data-Driven Imaging in Anisotropic Media 18 th World Conference on Non destructive Testing, 16- April 1, Durban, South Africa Data-Driven Iaging in Anisotropic Media Arno VOLKER 1 and Alan HUNTER 1 TNO Stieltjesweg 1, 6 AD, Delft, The Netherlands

More information

Symbolic Analysis as Universal Tool for Deriving Properties of Non-linear Algorithms Case study of EM Algorithm

Symbolic Analysis as Universal Tool for Deriving Properties of Non-linear Algorithms Case study of EM Algorithm Acta Polytechnica Hungarica Vol., No., 04 Sybolic Analysis as Universal Tool for Deriving Properties of Non-linear Algoriths Case study of EM Algorith Vladiir Mladenović, Miroslav Lutovac, Dana Porrat

More information

Physics 2107 Oscillations using Springs Experiment 2

Physics 2107 Oscillations using Springs Experiment 2 PY07 Oscillations using Springs Experient Physics 07 Oscillations using Springs Experient Prelab Read the following bacground/setup and ensure you are failiar with the concepts and theory required for

More information

Accuracy of the Scaling Law for Experimental Natural Frequencies of Rectangular Thin Plates

Accuracy of the Scaling Law for Experimental Natural Frequencies of Rectangular Thin Plates The 9th Conference of Mechanical Engineering Network of Thailand 9- October 005, Phuket, Thailand Accuracy of the caling Law for Experiental Natural Frequencies of Rectangular Thin Plates Anawat Na songkhla

More information

Experiment 2: Hooke s Law

Experiment 2: Hooke s Law COMSATS Institute of Inforation Technology, Islaabad Capus PHYS-108 Experient 2: Hooke s Law Hooke s Law is a physical principle that states that a spring stretched (extended) or copressed by soe distance

More information

e-companion ONLY AVAILABLE IN ELECTRONIC FORM

e-companion ONLY AVAILABLE IN ELECTRONIC FORM OPERATIONS RESEARCH doi 10.1287/opre.1070.0427ec pp. ec1 ec5 e-copanion ONLY AVAILABLE IN ELECTRONIC FORM infors 07 INFORMS Electronic Copanion A Learning Approach for Interactive Marketing to a Custoer

More information

Polygonal Designs: Existence and Construction

Polygonal Designs: Existence and Construction Polygonal Designs: Existence and Construction John Hegean Departent of Matheatics, Stanford University, Stanford, CA 9405 Jeff Langford Departent of Matheatics, Drake University, Des Moines, IA 5011 G

More information

ANALYTICAL INVESTIGATION AND PARAMETRIC STUDY OF LATERAL IMPACT BEHAVIOR OF PRESSURIZED PIPELINES AND INFLUENCE OF INTERNAL PRESSURE

ANALYTICAL INVESTIGATION AND PARAMETRIC STUDY OF LATERAL IMPACT BEHAVIOR OF PRESSURIZED PIPELINES AND INFLUENCE OF INTERNAL PRESSURE DRAFT Proceedings of the ASME 014 International Mechanical Engineering Congress & Exposition IMECE014 Noveber 14-0, 014, Montreal, Quebec, Canada IMECE014-36371 ANALYTICAL INVESTIGATION AND PARAMETRIC

More information

PHY 171. Lecture 14. (February 16, 2012)

PHY 171. Lecture 14. (February 16, 2012) PHY 171 Lecture 14 (February 16, 212) In the last lecture, we looked at a quantitative connection between acroscopic and icroscopic quantities by deriving an expression for pressure based on the assuptions

More information

Chaotic Coupled Map Lattices

Chaotic Coupled Map Lattices Chaotic Coupled Map Lattices Author: Dustin Keys Advisors: Dr. Robert Indik, Dr. Kevin Lin 1 Introduction When a syste of chaotic aps is coupled in a way that allows the to share inforation about each

More information

Analysis of ground vibration transmission in high precision equipment by Frequency Based Substructuring

Analysis of ground vibration transmission in high precision equipment by Frequency Based Substructuring Analysis of ground vibration transission in high precision equipent by Frequency Based Substructuring G. van Schothorst 1, M.A. Boogaard 2, G.W. van der Poel 1, D.J. Rixen 2 1 Philips Innovation Services,

More information

Testing the lag length of vector autoregressive models: A power comparison between portmanteau and Lagrange multiplier tests

Testing the lag length of vector autoregressive models: A power comparison between portmanteau and Lagrange multiplier tests Working Papers 2017-03 Testing the lag length of vector autoregressive odels: A power coparison between portanteau and Lagrange ultiplier tests Raja Ben Hajria National Engineering School, University of

More information

Poisson processes and their properties

Poisson processes and their properties Poisson processes and their properties Poisson processes. collection {N(t) : t [, )} of rando variable indexed by tie t is called a continuous-tie stochastic process, Furtherore, we call N(t) a Poisson

More information

AN IMPROVED FUZZY TIME SERIES THEORY WITH APPLICATIONS IN THE SHANGHAI CONTAINERIZED FREIGHT INDEX

AN IMPROVED FUZZY TIME SERIES THEORY WITH APPLICATIONS IN THE SHANGHAI CONTAINERIZED FREIGHT INDEX Journal of Marine Science and Technology, Vol 5, No, pp 393-398 (01) 393 DOI: 106119/JMST-01-0313-1 AN IMPROVED FUZZY TIME SERIES THEORY WITH APPLICATIONS IN THE SHANGHAI CONTAINERIZED FREIGHT INDEX Ming-Tao

More information

Easy Evaluation Method of Self-Compactability of Self-Compacting Concrete

Easy Evaluation Method of Self-Compactability of Self-Compacting Concrete Easy Evaluation Method of Self-Copactability of Self-Copacting Concrete Masanori Maruoka 1 Hiroi Fujiwara 2 Erika Ogura 3 Nobu Watanabe 4 T 11 ABSTRACT The use of self-copacting concrete (SCC) in construction

More information

Support Vector Machine Classification of Uncertain and Imbalanced data using Robust Optimization

Support Vector Machine Classification of Uncertain and Imbalanced data using Robust Optimization Recent Researches in Coputer Science Support Vector Machine Classification of Uncertain and Ibalanced data using Robust Optiization RAGHAV PAT, THEODORE B. TRAFALIS, KASH BARKER School of Industrial Engineering

More information

Analysis of Impulsive Natural Phenomena through Finite Difference Methods A MATLAB Computational Project-Based Learning

Analysis of Impulsive Natural Phenomena through Finite Difference Methods A MATLAB Computational Project-Based Learning Analysis of Ipulsive Natural Phenoena through Finite Difference Methods A MATLAB Coputational Project-Based Learning Nicholas Kuia, Christopher Chariah, Mechatronics Engineering, Vaughn College of Aeronautics

More information

TOURIST ARRIVALS AND ECONOMIC GROWTH IN SARAWAK

TOURIST ARRIVALS AND ECONOMIC GROWTH IN SARAWAK MPRA Munich Personal RePEc Archive TOURIST ARRIVALS AND ECONOMIC GROWTH IN SARAWAK Evan Lau and Swee-Ling Oh and Sing-Sing Hu Universiti Malaysia Sarawak, Universiti Malaysia Sarawak, Universiti Malaysia

More information

Figure 1: Equivalent electric (RC) circuit of a neurons membrane

Figure 1: Equivalent electric (RC) circuit of a neurons membrane Exercise: Leaky integrate and fire odel of neural spike generation This exercise investigates a siplified odel of how neurons spike in response to current inputs, one of the ost fundaental properties of

More information

Using a De-Convolution Window for Operating Modal Analysis

Using a De-Convolution Window for Operating Modal Analysis Using a De-Convolution Window for Operating Modal Analysis Brian Schwarz Vibrant Technology, Inc. Scotts Valley, CA Mark Richardson Vibrant Technology, Inc. Scotts Valley, CA Abstract Operating Modal Analysis

More information

UNCERTAINTIES IN THE APPLICATION OF ATMOSPHERIC AND ALTITUDE CORRECTIONS AS RECOMMENDED IN IEC STANDARDS

UNCERTAINTIES IN THE APPLICATION OF ATMOSPHERIC AND ALTITUDE CORRECTIONS AS RECOMMENDED IN IEC STANDARDS Paper Published on the16th International Syposiu on High Voltage Engineering, Cape Town, South Africa, 2009 UNCERTAINTIES IN THE APPLICATION OF ATMOSPHERIC AND ALTITUDE CORRECTIONS AS RECOMMENDED IN IEC

More information

The Wilson Model of Cortical Neurons Richard B. Wells

The Wilson Model of Cortical Neurons Richard B. Wells The Wilson Model of Cortical Neurons Richard B. Wells I. Refineents on the odgkin-uxley Model The years since odgkin s and uxley s pioneering work have produced a nuber of derivative odgkin-uxley-like

More information

Distributed Subgradient Methods for Multi-agent Optimization

Distributed Subgradient Methods for Multi-agent Optimization 1 Distributed Subgradient Methods for Multi-agent Optiization Angelia Nedić and Asuan Ozdaglar October 29, 2007 Abstract We study a distributed coputation odel for optiizing a su of convex objective functions

More information

Optimum Value of Poverty Measure Using Inverse Optimization Programming Problem

Optimum Value of Poverty Measure Using Inverse Optimization Programming Problem International Journal of Conteporary Matheatical Sciences Vol. 14, 2019, no. 1, 31-42 HIKARI Ltd, www.-hikari.co https://doi.org/10.12988/ijcs.2019.914 Optiu Value of Poverty Measure Using Inverse Optiization

More information

BUILDING TAGGING CRITERIA BASED ON AFTERSHOCK PSHA

BUILDING TAGGING CRITERIA BASED ON AFTERSHOCK PSHA 13 th World Conference on Earthquake Engineering Vancouver, B.C., Canada August 1-6, 004 Paper No. 383 BUILDING TAGGING CRITERIA BASED ON AFTERSHOCK PSHA Gee Liek YEO 1, C. Allin CORNELL 1 SUMMARY The

More information