xt76q52f8q86 https://exploreuk.uky.edu/dips/xt76q52f8q86/data/mets.xml   Agricultural Experiment Station, Department of Agricultural Economics, University of Kentucky 1972 journals kaes_research_rprts_13 English University of Kentucky Contact the Special Collections Research Center for information regarding rights and use of this collection. Kentucky Agricultural Experiment Station Research Report 1 : August 1972 text Research Report 1 : August 1972 1972 2014 true xt76q52f8q86 section xt76q52f8q86      
A   PERFECTING METHODS FOR PREDICTING THE COURSE
y OF RURAL AREA DEVELOPMENT
PART II
Forecasting Income for Selected
Rural Areas in Kentucky
By
_ Thomas H. Klindt, Garnett L. Bradford and Bruce R. Beattie
RESEARCH REPORT 13 2 August l972
L University of Kentucky : : College of Agriculture
 i Agricultural Experiment Station : : Department of Agricultural Economics
3 Lexington

 J x
x * a
’

 i PR IC I·`,»\ ( LIZ
    IN lII(` $(`(`(lII(l wl I\\\) ll'l)<>l`l?» I'l'I.lIIII; l`(`\lIll\ (ll Ir('§('QlIr('I\ (r'(lII(IlI(`Il`(I IIIl(l(`I` I{t'|llllt`l<}'
. Il.ttt‘h l’1‘<>_i¤‘t‘t H9, "l)<·\<·l¤»pnit·nt til I’r<><·t·iltirt·> hu (lgimiitiltiiig .in(l .·\>»<·s»i11; tht- l‘.u»1t·»tlii<
 · \\4t·ll·ll<‘lltg wl I{ttl.tl .`\I('tl\ in I\It‘lllt|t‘li}‘,U II1t· rt·>t·.iinIi r¤·I>·»rt<·¢I in thi> .ti1¢l its t‘¤»i*tpt¤l1l<>11
  ]>ul>lit‘.tti<>ti is l>.i>t·tl in l.ut;t· l).Il'I tm tht· l’h.I). r<·>t·.iit·li til 'll1<»1ii.t>II.Iiliiitltjlliix tli<:»i» rt·>t·;i1t’li /
_A  rt·p1‘t·>t·11ta·¢lthci11iti.nlctiiitiilitntiniitt»II.itL1t`><; ttl ltllkll MTA
 I t·t·<>iit»iiii¤‘ nlcwltrpiiicint. Spt·t·ilit·;ill5, lite t»h_jt·t·tixt·> iirq imtilwrl (ly tIt·liiiiti`·ii wl cc<>¤i<¤iitic`
 I (lt‘\¢‘l¤>}Hll<‘l1l. (2) tl<‘lillt‘.1li¤»1i <»l ·jiltt‘|‘i.i .Ill(I Il1’t>trt1t‘ti<»i1 till .lII('I`II.Ill\<` m¤»l`(ll(`Illlg iiiiptntttiit c<»ir1pt»iiciit> nl cum<>iiiit`
5 tlt·xt·l¤»piiit·i1t, (·l) ¢t»iiip.u‘.itixt· lt‘>ls ··l .•ltt·1‘11.¤ti\ c mtwclcls iii ;1t`ui1`d;11icc with c1‘itci‘i.t c>t;1bIishcd
  in t»l>_it·t·tixt· txx·t»_ uml [Bltt><‘¤·l"°l1t·~t"iiititlvlsl>.t>c(I<»11i’t·>L1lt><>l>t;ti1tc(l in carrying out objective
 i lt»tu‘ to pi‘t·elit`t tlit· l`l*lll`$l' ·»l t·t‘¤»i1¤»iiiit‘ tlvxcltiiiiiiciit hir >clt·t`tt·cl tttiul ;ii‘t·;i> til Iiciitucky.
 I ll`hl> }>t1l>ll<‘.ili<>11 l>1‘<·¤t·1it¤ t·mpii‘itiil insults l>t·i`t_iiiiiii; to <»l>_it·gti\‘c< throw. l<>ut‘. md l.l\‘C.
 I Spt‘<`illt‘.nll), >inp;lt‘-t‘t]t1.tti··1i iiittimc 1·»1‘<·t`.t»ti|i; mt>tlt·l~ .ii‘t· t·<»|i»ti‘ttt‘tt·tl, tt·>tt·d. ttlttl U>CCI N1
 i pivtlivt t·»t.tl .u1tl pm t‘.tpit.tl [)t'l`\<>llttl iiitwiiiv in »t·l¤·t·tt·tl i‘ti1‘.il.ti1·.i»til Iit·ntut`lvcrx‘icxx‘ ul
  the t<»t.¤l tt·~t·.tu·|i ¤‘l'l<>1t .1 tlulinitittii wl (`(`i>IltlIl1lt`(It'\L'I|*})l]lL'III imtlt‘i`itci‘i;1li<>1`L‘\;1lL1Ltli11glTl<>d€l
 I t‘lllt’.lt`§' is t`<‘p<>t‘lt‘(l iii .¤ t‘t·iiip.u1i·>11 pit·t‘t· tt: thi> !'t`})t\I'IZ I'. rn tiziiilg _lIt·!/zodx _/or [)7‘<’(IZ4<`[1i7I_g’[/Z<’
I (.`t2Hf.\<’ U] [fum! rl/“¢’(l I)»·z·t*/ti/miryii; [Girl [ V/Qtzyiizqf ii ]}¢'~/Hllvtllvttll Ur [;`t·mzm2iz-{ I)t‘z‘t‘/rijmivili (UIC!
 i` tl /"l‘<1rl1<‘i<‘~f/6 jnr I'.-Z`(l/I((lil[i71`Q ,\I¤»t[t·/ I;i[_]it‘i1r·x· (I{t·>t·.i1’cli I{t·}>t>1‘t ll, L'iii\t·i’>it} wl I{n‘\1ll1Cl<}"
· l)t·p.u‘tmt·iit ul .~\g1‘it’t1lttt1‘.il I·.t·<·zi<»iiiit`>. HITS).
 · llhv .lllIlI\il`S ;[l`.IIL`lIlll}` .it·I4iitmlt·tI;;t· tlit- .t»>i>t.itit`t· Anil tw>·>pc1‘;iti<»ii nl I)i‘. II;l1`Ul(I Ii.
I (ll1.tl‘I<‘~\\’<·lllt .tlt(l sl.1lil ol {liv ()llit‘t‘ til I)t·\t·lt¤}»mt·iit St'I`\IL`l`\ .11itl BL1>IlIL`>> R¢.‘>C.tI’CIl All lllf
I l`nixt·i‘>ity nl Kcittticky, Mi. \\'iIli.tm (Q. Ilcizcl hid sttill tat tho I)I`t>gl`.1I`ll tmd Rcscttrcli Stull.
 * Iin‘lllll<`l·i) I)<‘pau‘l1m‘ttl til I{c\t·iit1t·_ .mtl tht- Iiuiittigkt I)<_'l).ll`II]`|L'I1I t>li Ifctittwmic Sccllflly li<>l`
_ pmxitliiig muyh t»l` thu tl.it.1 lor this >ttidy. .·\lst·_ the ci‘itit`i>ms .iticl Utilltltltflils ul- our eullctigttcs in
;_ t|it·1)t·p.u·tmt·nt ol ,·\gl`lt`llllll1`;1I I·Qtx»iit¤iiiics .lI`&` gixitciittlly .ipprcci,itctI.
 L
  3
¤ l

 (]()N'1`1·LN'1`S
Page
1’R1·Q1·`.·\C1·l .......................................... 3 /
1.181 ()1·` '1`.·\1i1.1·QS ...................................... 6
1.151 ()1·` 1·`1(}L`1{ES ..................................... 6
~ OB_]EC1`1\`ES ........................................ 7
I\1()1)1~.1.S:(}1iN1·lR.\1,1)1S(Z1'SS1()N ............................ 8
SIMPLE 1·`()RE(ZAS'1 X1()1)1·Q1.S ............................... 11
POLICY-OR11·LN'1`1·Q1) 1·`()R1·Q(IAS'l`N1()1)EI.S ........................ 13
· CO}\11’.-\1{.\11\`1·, \1()1)1·Q1, 1`1·)S41`1N(j ............................ 16
1971 P()RECAS'1`S ..................................... 21
SL`31.\1ARY ANI) CONCLUSIONS ............................. 22
APPENDIX A ........................................ 25
APPENDIX B ......,................................. 26
5
l

 LIS'1` OF TABLES
Table Page
1. Types of Income Forecast Models Studied ....................... 8
2. Selected Regression Results, Lagged Income Model .................. I 3
3. Selected Regression Results, Economic Sector Model ................. 15
y 4. Predictive Accuracy Statistics, Per Capita Personal Income, in Dollars ......... 19
I » 5. Predictive Accuracy Statistics, Total Personal Income, in Dollars ............ QU
A 6. Predicted Total and Per Capita Personal Income for 1971, in Dollars .......... Q}
A 7. Predicted Total and Per Capita Income for Selected Kentucky Area 1)evelopment
Districts, 1971, in Dollars ................................ 2]
LIST OF APPENDIX B TABLES
A I. Observed and Predicted Levels of Area Income, Income Extrapolation Model. 1968, in
Dollars .......... . .............................. 26
2. Observed and Predicted Levels of Area Income, Lagged Income Model, 1968, in
Dollars ......................................... 27
_ 3. Observed and Predicted Levels of Area Income, Economic Sector, Resource·Based
Model, 1968, in Dollars ................................. 28
LIST OF FIGURES
Figure
1. Development Districts and Areas ............................ 10
I 2. Components of Model Efficacy ............................. 17 »
l >
/ I  

 l’l·]Rl·`l·L(I'l`lN(1 I\ll£'l`IlUl)S l·`()R l’RIil)l(j'l`lN(} lllli COURSIC
()l·` RURAL ARICA l)IjVliL()l’l\IlQNl` l A
Part ll
 
l·`()Rl·L(Z.·\S'1`lN(} lN(J()l\ll·Q l·`()R Sl£I.l£Cl`l·QI) RURAL
.»\Rl£AS IN KIiN'l`U(§KY
hy
»
ll`ln»ni.is ll. lilindt, (Lirnett L. lirttdliord and Bruce R. Bez1ttie*
t
Rtnxtl tlt·x<·lt»piiieiit iese.nt·li is eert;iinl3 .1cti<>n. Lsle [5, pp. T8] ewntended that
nntlnng new_ yet there li.is heen gi new surge minimum needs lor exuluutiuii include:
wl iltlt‘t`t‘st tit t‘t‘t‘l;titt .tst1t‘t‘Is¤vl tltls tt‘st‘.tl‘t`lt,
espetitlly sniee IW}?. Ihis interest w;is lhe ltttttre euiitlitiwtt expected to prevail
exliilntetl in the lintlings wl the l'resitlent`s in the gthsenee <>l` the plttnnecl ellects ol
N.tti··n.il .\tlxis•¤r) (I·nnniissi·>n nn Rtnxtl new prt»gr.nns or prujeets pr<>p<>se(l. .·\
l’t»xeitx_ where it w.ts pnintetl unt th.tt rtnxtl eurxe clrgiwn to describe this condition
st·t·tt»is wl the l`niterm;1l"
tlispinpniti·>n.ite sIi.nie nl the n.iti¤»n`s ptwert) growth p.ith. lhe nnrnitil growth p;ith...
|¢i|. |·tn‘thei, the l’1esitlent`s l.isk l·`t»ree tin reliers tn expeetetl lluture gruwtli in the
l{nr.il l)t·xelt¤pinent iiitlitqitetl th.it eetinninie .ihsenee nl newly planned nr un.intiei·
i prnhleins in rtnxil .ne.is (ln not l`L'l11.llll puted prwgrtnns, projects, teeltnultigiettl
twiilitietl tw rtnxtl .n‘e;ts. l)is.1tl\’;ntt;1germ;itinn, together with
.in .inswer inginy to lintl jobs, but ;t expected elleets lrom tt planned program or
tlispinptirti·»n.ite ntnnher tn line wellttre and project, yields the inlormtititm required to
g sltnn hnnsing." [7, p, —l|. determine the ;1etu;1l impact ul ti proposed
(Innsistent with the .iht>x’e-inentinned program orprtwject.
st.inee, pnhlie policy in.ikers h.n·e exhihited;1
renewetl interest in rnr.tl development; Objectives
ltrin.iti<»ti tn ex‘;iln;tte .tltern;ttixe enurseseil .—\n meiwiew ol some resetireh tilting
these (ill)U\`L'·ll]L`llllUIlL`(.l) lines recently
____ ____ _ ,_,___,___V etindneted nt lientneky |—l| is presentetl in
J this report. 'l`he prnnury einphitsis was npnn
Wlssistnnt l’mI`essnr of .·\grieuItnr.il lieminmies, Luuisinnu g(m$[]Ug;iug gimp];-_ yet Ql`lt‘([i\’Q mntlels to
7 L     t‘fi*T·‘*; ·*·‘ F ·‘··   <·‘i ··‘¤r*.€¤'*·*i·' t·t,,·mst Wi mi pt—tt.i,.iit. pt-twin i.t.—.)mt— »
Acntitviiiits .intl .·\sstst.1nt lmlnssur ol Agrteultttnil
lieonornies, University of Kentutrky, respectively. l`Ol` SClCClL`(l l`Lll`Lll ;ll`CLlS ol. KL‘lllLlL`k}' lll1(lCl` Il1L‘
7
l
 

 8
assumption of Lyle`s "normal growth path."l expressed assertion, especially among rural
Resulting forecasts (predictions) may be used development decision makers, that
by policy makers to evaluate the impact of economists have achieved a level of model
l proposed progrmrrs gr prgjqgtg {O determine building sophistication that far surpasses the
which, if any, should be implemented. capability of action groups to apply such
Further, such forecasts may give clues models in the solution of their problems.
concerning which rural areas will be in More important, however, was the practical
greatest need of public assistance if no action expediency of first working with relatively
is taken, thereby allowing a degree of simple models, then hopefully moving on to
forewarning to policy makers. more complex ones such as simultaneous
_ Specifically, objectives of the study equation systems, simulation, etc. liconomic
_ ( were: and statistical theory underlying
single-equation systems is much better
Q (l) to define economic development, developed; and such systems readily lend
( (2) to delineate operational criteria and themselves to the "proven" statistical
_ procedures for evaluating model
efficacy,
( (3) to construct alternative models for
» predicting important components
t (levels Of variables) gf eggrigmig rl`3.lDl.€ 1.-—'l`ypes of Income llOI`CC(lSC
A development in selected rural areas, llOd€1$ StUdl€ § R
  ‘
  Q P9   `
;    _i ;    
I IQ‘*¥;i?iE¢¥¢E ;¢E¢2¤-1;¤;1; -.·.‘ ‘ IA
Y w Q 4`?***°·-"I?‘?¥?` II.f§ZYYI "
    ‘*?I?¥?¥$i?*¥=*I>¢¢2%>,  riTI;t;  
I ;;IQ;iig;¤€2*;>>Qijizrq -%2€;€2§;2; IA
I I Ii€i5{i{i; `T;??I€I¥{? }:€i »:?;·1·;-I *5
` v f;Q;Q§iQ;?Q ‘_.A `;Q?1Q;¥;? :Q;Q;éQ:QiQ;Q;Qi  
ii".-t` i1?;` ‘  I ·n I
I I I I II nl
 III Ig I§I' z
I`I§:r_l1  
. I I 1 ’
  W  ¤
  I I 'I·°°§'I ` 2
2 I IIIIIIIIII
;•••••q 6•¢ .•'•• II I \ ,_]
»:·:¢z¤:’»5==:¢E=§¤*°¢ I I I · »
c . .•$•Z•°•$4é?•§Q     ’ —.   I
I ;¢•¤•$:.•·•·•     
I •,•:::?§b$g‘   _'  f_       j   ,
T, ...·°•¢;.•.¢v$•~I  54- II.-     —...I.»» I
I I C iw     V._.    
Fi II 2 a =¢:$•’€'.   ‘*        = I
> ,-4 '••V _ I I          
i 3 2 is $2 ;* ~•’·1é$g§!;2      -- » II; II_, ’
> GJ ·r—l ,¤ PI 5: •¢•O··   I   ‘ .
.,.4 ,..4 D; E 3 ’•••°=‘\•¢   - 4.»» I ,. .I I  
1 cd ·r-1 :5 Q O 9•?•¢   _   ‘;   `
5.. ;:; LJ ,-4 ;¤ __.» -   ._·,.—
{Z >~, (D cd I   ~ Z
I no ;: :-4 w ¤+-4 cu       I·.· A
» » ¤> ¤¤ <¤ ¤ :4 I .·_. ·
LD an cc A ¤¤ <:   ·‘
@ I.I’f<~ IIII E   I
I I •-
I I i __

 I I
ttsclulnoss. ln this roport, only tho lirst ol tltrtro·yo;trprodiction intcryttl." 'll'1U>,llIl'llL'Ll
thcso two typos ol tnotlols tho oconornio it wtts ttssumotl tltttt I965 Wtts tho Ittst your lor
sottor ntotlcls is tlisoussotl. In thoso moclols which incomo wtts knownf Pttrttmotor ,
.ttItIition;tl policy inlortntttion is obt.tinod by ostiintttos (Cth ttnd dl) lor ottch ttrott woro thon
ltypotltosizing .tro;t incoino to bo at Iittnction ol usod to tioroottst incomc in tho most rocont
tt-tctgtl intlcpontlcttt yttri.tbIos roprosonting your lor which incomo obscrxtttions ttctuttlly /
ctontttnic .tctixitics in xgtrious soctors ol thc wcrc ttyttilttblo in this ottso. I968.
ctottotny. 'lhc ntotlcl yioltlotl inlortntttion (Zonsc<)uontly tho tcst-Iorotgtst otitttttion nitty
_ tontctnittg tltc rcI.tti\o inconto cllocts ol bo writton tts
t.triotts I>r<>.ttI oconoinic scotoi s. lhoso sootors _
wcrc rcptcscntotl by physic.tI ttctiyity \.tri.tbIos XYGH Z on + (it) I (LII
sttt It .ts otnploy inont .tntI .tcros hitrxostctl.
whoro
Sitnplo l·`oroo;tst Nlodols
ctw Z uw Ct) Z 6), ttntl II` Z IU.
— |.ntc.tt cstr.tpoI.ttion is prob.tbIy tltc
lllttsl toininon tochnitluo tisotl to Iorcottst 'Iho sttporscript nottttion t` I is ornployod to
t·tonoinic x.tri.tbIcs. lho cxtr.tpoI.ttion tnotlol idontily Iorocttstod x.tluos ol tho pttrttmotors,
txl1ichxx.tsIittctIn1.tybospooiliod.ts ctw .tnd ul. In subsotiuont discussion, tin
oxpl.tn.ttion is .trk intolving longor prodiotion intorutls would Iuxobcon
·t\`.tIl.tl)l&‘   llI62_ ;tlt(I   to })L`t`l]tll Ll Closiratblo, hut cl.tt;t rostrictions limitod prctlictions to throo
tost ol pt·t—tlit:tit·t· tttctttntcv, ttssuming tt “"‘Ii" T"l""i"“ *`.""II°`I""T_`Y'f`_“' mm" bm "`lmi "I" Hill
· · prosontud hcro. I\t»to, Itowtxtt. th.1t till modtls ttncl test
prttooduros nrc constrtiotcd such th.it tin .t\‘iously. it would liuvo boon tlosirttblo to Iittvo moro
oomploto tiinc scrios tluttt. llowoxcr. it is tho .tuthots`
_________¢_________ oontcntiun thttt in Iiormultttivo roso.troh ul this typo it is
bottcr to hzuo rosoatrch rosults which h.txo boon constrttinod
I G by luck t¤l`d;1t.1 than it rosottrch projcct which is still dwttitittg
_ ` S<·<·_l•thttsttm [3, pp, 3-1 1 |. "c°d)~ And   d' lnOdLl institutions and   value of breeding livestock. They were
having five independent variables which also statistically urmsrgrrrrrtarrx at the 0.05 level.
t
.. -*'*<.,.‘,"\‘

 14
This model was fitted cross-sectionally for Forecasts for X3 and X4 (X3 and X4)
each of five years—1959, 1962, 1965, 1966 were based on a fit of the following
_ and 1967. §ome of the regression exponential model for each of the 18 areas:
estimates·R2, B and t values—»are shown in
Table 3. Note, that practically all of the X: G tltte (7)
4 variation in total personal incoinc could be
attributed to variation in the four where
independent variables and that a sizable but
lower proportion ofthe variation in percapita X = X3 or X4 in years 1954 through
personal income could be similarly attributed 1965;
» to the independent variables. This was true
— B for all years. Presumably, the difference (in t = an index of the year, i.e., 1954 I 1.
R2 values) was because per capita income is a 1955 = 2 ..., 1965 I 12;
{ ratio of two variables (total income and
_ population) and, in this case, was subject to ® and tit Z model parameters;and
5 much more variation. Lower t values for the
respective regression coefficients of the per e = the error term.
T capita results may be similarly explained.
l)erivation of the test-forecast equation listimates for tl and tl; were then "pluggcd
_ involved forecasting new levels (values) for into" an equation ol the same form as (7).
V the independent variables and for the and forecasts of X3 and X4 for 1968 (X3 and
i parameters. ln most other respects, the X4) were obtained for each area by setting t ¤>
V methodology was identical to that employed 15.
. B for the simple forecast models. Assuming Trend analyses also were made for the
1965 was the last year for which data other two independent variables, X4 and X2.
. (observations for Y and Xl through X4) were However, no significant trends could be
available and that a three—year income established for X2 (croplandacresharvested).
forecast was to be made (into 1968), the lt was decided that even though values for Xl
test-forecast equation could be specified as (burlcy tobacco acreage) have been tending
4 4 4 4 4 4 4 4 4 4 downward they are so subject to decisions
Y68 Z BO + BIXI + B2X2 + BBX3 + B4X4 (6) made vis’-a·vis’ the price-support prograin, it
would be presumptous to forecast this
where the notation (coefficients and variable. ln short, it was assumed that values
variables) has similar economic meaning as for for Xl and X2 would remain at the last
, the stochastic form of thc model Spscificd in "1 (1.25) (1.27) (1.91) (2.08)
* A
.1 B3 0.06333 0.05486 0.02623 0.01871
H (2.95) (1.69) (0.75) (0.50)
1 84 0.05850 0.04544 0.06537 0.07069
(1.93) (1.24) (1.49) (1.66)
 
aTotal personal income is estimated in thousands of dollars, per capita
M income in actual dollars. Calculated t values are shown in parentheses
8: directly below each B value (13 degrees of freedom). i
if
Iii
lu

 16
_ Forecasts ofthe coefficients (B0 through Similarly, four sets of parameter forecasts
j B4) were derived using each of the following were derived for total personal income.
` four decision-rules (model variations):1 3 Altogether, 54 test—forecast equations
were formulated~(a) one time extrapolation
( Variation A——se1ect value from most recent year equation for each of 18 areas, (b) one lagged
(196519 income equation, and (c) four each of the
. . resource-based and investment-based
. Variation B-analyze the array of estimates _ . ___t d l_ h_r_ _, _h { ,)_ t~
_ (e.g., Bis) for a trend and if a significant trend was tconomlc bu' or mo C S' W L L LAL 11* _U
i detected, extrapolate to [hC {CSI-fOl'CCa$[ YCKI     [(u)’   und   `vus dcvclopcd ijor
· lf no significant trend was detacrad_ the value {Oi- the both per capita and total personal income lor
‘ ***051 f¤¢¢¤i vear(1965) wasused. the three-year prediction interval. With the
. . . use of these test-forecast equations, personal
_ Varxatton C-same as Variation B, except if no - { {,1 _ 1 _r _, T) ,_ _ [ WC, (Cd
; trend was detected, then a simple average of the array income ( O il imc pL fdpi A WM U fb
O{B·Swas uScd_ int