1930s Economic Recovery

Question 1

Features of the Recovery (8) AMSRGM

Answer 1

• Depression lasted from Q3 1929 to Q3 1932 (real GDP down 5%, industrial production 11%)
• Recovery unusually long & strong – real GDP grew 4% p.a. 1932-37 (unmatched, historically)
• Alford (1972) – “Depression and Recovery?”
  o Long term economic growth rates comparable 1924-29 and 1929-37
  o Unemployment problem remained
  o ‘Difficult to see how one can speak of “recovery”’
• Mitchell, Solomou & Weale (2012) – estimate monthly GDP for GB, 1920-38
  o Use set of monthly indicators constructed by The Economist (14 index numbers)
  o Trough is September 1932, then recovery begins in October 1932
• International comparison (Solomou, 1996) – GB did very well
  o Most countries witnessed retarded growth in 30s (inc. US, France, Italy & Japan)
  o By contrast, GB grew on steady trend path since 1920s (2% p.a.) & saw improvement ct 1913-29 or 1899-1929 (+ saw rapid fall in cyclical u/e rates from 1932 peak)
  o GB cyclical performance from trough to peak is at median of distribution, but strength of recovery related to severity of depression – UK depression was comparatively mild, so median position is pretty strong (those who do better had much deeper Depression amplitudes)
• Reinhart & Rogoff (2010)
  o Examine 21 “now-advanced” economies’ performance during Interwar period
  o Statistically significant difference in international growth rates 1919-28 & 1930-39
  o Hence most countries find it difficult to recover and see reduction in trend growth
  o In this context, GB performs very well
• Greasley & Oxley (1996) – perform econometric tests on GDP/industrial production to see if paths are trend stationary – find that 1929 crash had no more than transitory effect in GB
• Expenditure patterns during recovery
  o Consumption expenditure was very important through 1932-37
  o Export demand/domestic investment more important in earlier stages (1933-35)
  o Government expenditure important in later stages of boom (rearmament – 1932-37)
  o Recovery was internally generated – export volumes in 1937 below 1929 levels
  o C & I accounted for 75% of the recovery (Morys, 2014)

Question 2

Natural recovery intro (4)

Answer 2

• Richardson (1967) – ‘the role of Government Policy, if it were positive, was minor’
• “Natural” Recovery idea
  o Market economies adjust to shocks via price/wage flexibility
  o Business cycles are short-run deviations from trend
• ‘Recovery that occurs without any conscious effort by government to affect through policy’
• Can occur via both demand or supply side
  o Demand side – Keynes effect, Pigou effect (but if net debtor, then negative) & RER effect (see ToT, below)
  o Supply side – technology diffusion, shift to new industries & real wage moderation

Question 3

Natural Recovery - New Development Block - FOR view - Aldcroft & Richardson (1967)

Answer 3

Aldcroft & Richardson (1967)
* Expresses view that strong growth path in 1930s is due to new industries (chemicals, electricity, electrical engineering, motor vehicles & consumer durables)
o New industries form own “new development block” with strong input-output links within the new sectors and high complementarity of demand between them
o Also not reliant on exports, so GB suffered less from weak overseas demand
o Indeed, they grew very strongly in the 1930s
* Structural transition to new, capital-intensive industries may provide explanation for high unemployment of 12-14% in early 1930s and subsequent decline

Question 4

Natural Recovery - New Development Block - AGAINST views (5) KVHCK

Answer 4

• Kitson & Solomou (1990) – very small impact due to small share of output – just 10% of manufacturing output in 1924 and only 20% by 1935
• Von Tunzelmann (1982)
  o Calculates a counterfactual, with new industry growth held at 0 (via “Leontief inverse” method and input-output tables) – output only 3 % lower in 1935 – economy growing at 4%, so new industries don’t explain recovery
  o Input-output tables from 1930/1935 Censuses of Production show that backward linkages ran predominantly back to old industries – hence, recovery is driven by both industry types (and policy may have a role via old industries)
  o Conducts shift-share analysis, decomposing productivity growth into “within sector” and “structural change” parts – former more significant (and even across old/new)
• Hatton (1988) – doesn’t contribute to demand at all
  o Exports – no, because mostly import substituting
  o Investment – no – only 7% of total net I (32-37) (most of which in housebuilding)
• Crafts (2018) – manufacturing sector was transformed in decline in old staples in favour of new industries – to be expected in any dynamic economy
  o New industries do not appear to have significant comparative advantage – old staples persist as UK’s strongest exports by end of 1930s
  o New industries are important
   Experienced faster productivity growth of 3.1% p.a. vs 1.3% p.a. for old
   Constituted 33.9% share of overall manufacturing productivity growth
  o But unrealistic that new industries transformed UK productivity performance
• Kitson & Michie (2014) – it is very hard to define old/new industries – is it just based on their growth rates? (so endogenously true)

Question 5

Natural Recovery - GPT - intro (3)

Answer 5

• Bresnahan & Trajtenberg (1996) – General Purpose Technologies (GPTs) are:
  1. Pervasive – GPTs spread to most sectors
  2. Improving – GPTs get better over time, so keep lowering the costs
  3. Innovation spawning – GPTs make it easier to produce new products/processes
• Examples include electricity (1894 to late 30s) (+ve impact on labour productivity in manufacturing) & ICT (early 70s to modern day)
• Jovanovic & Rousseau (2005) – evidence shows that GPTs arrive episodically, at first creating turbulence and lower growth, and then delivering higher growth/prosperity later (not true – productivity growth definitely occurs)

Question 6

Natural Recovery - GPTs - FOR arguments (3)

Answer 6

• Evidence suggests that UK could have benefitted from GPTs in the 1930s
  o Electricity adoption well behind US (pioneer), especially in 1920s
  o Follower countries should be ready to adopt electricity in 1930s, creating boom
• David & Wright (2003) argue that UK did benefit from it
  o National Grid was created in UK 1929-33, so diffusion matched by US by end of 30s
  o GPTs are capital-saving technological change, so should see lower K/Y ratio – indeed, Matthews et al. (1982) find K/Y ratio falls in all manufacturing industries 1924-37
  o However, TFP growth still a lot less than US rates in 1920s
• Jovanovic & Rousseau (2005) – Hodrick-Prescott (H-P) filter suggests productivity pickup in latter part of electrification era

Question 7

Natural Recovery - GPTs - AGAINST arguments (2 + 2)

Answer 7

• Kitson & Solomou (1990) – evidence suggests there was some technological catch-up
  o Britain closed labour productivity gap to US for first time in 50 years over 1929-38 – labour productivity may proxy for technological gap
  o No reason why it should close in 30s – sizeable in 1913 and 1924 too
  o Hence reason is favourable AD conditions – not just S-side in isolation – interaction
• Ristuccia & Solomou (2014) – question notion that widespread productivity increase resulted from electricity diffusion
  o Productivity bonus of the UK manufacturing sector can be observed over the entire interwar period – challenges the GPT theory since the UK was far behind the US in electricity per worker at the start of the 1920s
  o (David, 1991) GPT theory indicates that GPTs do not deliver productivity gains immediately upon arrival (time take for diffusion) – but productivity gains come about much earlier than implied by GPT theory (from early in 1920s)
  o Productivity growth paths of the UK, Japan, Germany and France differ significantly, despite sharing a common path of electricity diffusion – UK and Japan experienced trend acceleration across the whole IW period, whereas French productivity grew in the 1920s but stagnated in the 1930s, and Germany experienced no trend acceleration in productivity
  o France – implies that countries limited by policies (with France remaining on the gold standard in the 1930s) were unable to take advantage of the new technology linked to electricity – suggests strong and important interaction between GPTs and policies
   Risky comparison – electricity may not have been as effective in France due to higher proportion of workers still in agriculture vs UK – expensive to bring electricity to rural areas
• Supply-side is not enough – merely defines a potential – need demand to realise it
• Policy is important in this regard – interactions of policy/S-side are key to recovery

Question 8

Natural Recovery - Terms of Trade Effects (4)

Answer 8

• Terms of Trade ToT=Export Prices/Import Prices, so increase means can buy more M for each X & may also be lower cost-push inflation (import price inflation lower than export price rises)
• Collapse in commodity prices abroad meant ToT improved by 20% 1929-1931
• ToT improvement meant real consumption wages increased (lower prices), increasing C
• However, ToT deteriorated 10% over 1933-37, suggesting it was not key for recovery, just ameliorated the Depression itself

Question 9

Natural Recovery - Real Wages/Labour market (4) (BDHK)

Answer 9

• Beenstock et al. (1984) – real wages key to recovery
  o Higher product real wage reduced the demand for labour (especially in tradables), reducing profits, investment, output & employment
  o Reversed from 1932 as product wage growth moderated – led to growth
• Dimsdale (1984) – in econometric estimates for labour demand, significant negative coefficient on product real wage only obtained in presence of demand variable (e.g. world trade), suggesting that demand side factors are key
• Hatton (1988) – moderation probably just the normal response to Depression – movement along SRAS, not a shift (Keynesian view)
• Kitson & Solomou (1990) – flaw in other studies is use of wholesale prices, not the GDP deflator (using right measure, no reversal in W/P until 1936)

Question 10

Natural Recovery - Housing Boom - Why? and FOR arguments (4)

Answer 10

• Why? – 1. enhanced availability + affordability of mortgage finance, 2. permissive land-use planning rules 3. shortfall of investment in 1920s
• (Worswick, 1984) Housebuilding accounts for 17% of GDP growth and 30% of increase in employment between 1932 and 1934, including multipliers to related sectors
• ‘Cheap money hypothesis’ – decline in interest rates following July 1932 war loan conversion coincided with and was a major factor in causing housebuilding boom
  o Intuitively, as i/r decline, shift in savings towards investments offering fixed rate of return – PROPERTY
• Average monthly mortgage payments fell 9% 1931-33 – also some demand-side effects as housing became affordable for more families

Question 11

Natural Recovery - Housing Boom - AGAINST (6) FHKBW

Answer 11

• Lack of synchronisation between cheap money and housing boom
  o Residential construction growing in 1920s and experienced boom in 1927 comparable to – Feinstein (1965) puts forward weaker version of hypothesis, that lower interest rates sustained (rather than stimulated) the boom
  o Argument – building societies, experiencing strong inflow of funds, forced by law to lower mortgage rates and increase length of loan, thus acting as a transmission method for falling interest rates
• (Humphries, 1987) – effect not via cheap money
  o Rate of return on building society shares and consols had previously been comparable given their similar risk
  o War loan conversion did lower the returns on consols, creating a differential in favour of building societies
  o Would expect this to result in more rapid growth for building society funds – in reality, deposit growth faster in 1920s vs 1930s
• Kitson & Solomou (1990)
  o Important for turning point (17% of GDP increase 32-34) but not for 1930s in general (only 9% of overall GDP growth 1932-37, ct 50% for manufacturing)
• Broadberry (1986) – increase in building society advances in early 30s was not due to higher deposits, but due to lower liquidity ratio due to competitive environment & high growth of some societies (but still attributes ½ of housing boom to MP)
• Housing only 3% of GDP in 1932 (Worswick, 1984)
• Natural? Linked to monetary policy

Question 12

Natural Recovery - Conclusions (3)

Answer 12

• Worswick (1984) – ‘policy made important contributions’
• Kitson & Solomou (1990) – all reasons are empirically invalid/not capable of explaining growth by themselves – S conditions favourable, but only define potential – need policy too
• Housebuilding best (but role of MP), GPT only explains supply, NDB worst

Question 13

Devaluation (Intro) AEESR

Answer 13

Accominotti (2012)
o London merchant banks and acceptance houses were small, poorly diversified, more vulnerable to liquidity risks and engaged heavily in trade finance.
o When there was suspension of foreign exchange payment due to capital controls in defaulting countries, people feared that small banks might face serious liquidity and solvency risks, so started withdrawing their deposits.
o Since banks had little capital and reserves to absorb the loss and they were an integrated part of the whole banking system in the UK, the stress was transmitted to whole financial system
o BoE aimed to prevent systemic banking crisis to avoid losing its comparative advantage in financial services through London - defended the £ with interest rate increases and OMO but run continued
o BoE also had rapidly falling gold reserves, which would run out
o So devaluation only option left in Sept 1931, preventing a large crisis

• Eichengreen & Sachs (1985) – take comparative approach to devaluation – very positive
  o Divide countries into 2 groups – those that devalued & “Gold Bloc” (Belgium, France, Italy & Switzerland) – have a total of 10 countries
  o Propose 4 channels for devaluation to effect economic performance – find all 4 key
  o Graph % change in industrial production against ER level (1929-1935) – downward sloping, so clear relationship between higher depreciation and higher growth
• Eichengreen (1991) – figures show that depreciators/sterling bloc had a lower amplitude Depression (due to early action) and a more robust recovery
• Solomou (1996)
  o Estimates gti=a+bRegime+cAmplitude+i (Regime=1 for devaluers)
  o b is positive and statistically significant (t=5.57)
  o Hence, controlling for amplitude, average growth was 3% p.a. higher in devaluers
  o Concludes that initial +ve was trade, persistent +ve was freeing of MP
• Reinhart & Rogoff (2009) – cast doubt on positive effects of leaving Gold
  o Broader set of countries in sample means effect not visible (especially Latin America) – not that important – many have changed sample and found same result
  o Effect not visible if other measures of business cycle are considered – important criticism since most work is done with industrial production (easily accessible data)
  o Graph GDP and ER level (1929-35) showing no clear relationship (23 countries) – may depend on what type of country you are (debtor/creditor, MP use etc.)

Question 14

Devaluation recovery channels: Channel 1: Real Wages (NO) (3)

Answer 14

• By causing inflation, devaluation moderates real wages (assuming nominal wage rigidity)
• Eichengreen & Sachs (1985) show 2 graphs for this case
  o Negative relationship between W/P and industrial production
  o Positive relationship between W/P and exchange rate
• Madsen (2004) – the incorrect wage deflator is used – misleading
  o Eichengreen & Sachs (1985) use wholesale prices – inappropriate
  o Wholesale prices are misleading – commodities have unduly high weight (fell rapidly), includes M prices (related to ER) and agricultural prices (crisis in 1930s)
  o Regressions on wage-setting imply that wages not indexed to wholesale prices, but to value-added price deflator
  o Using correct deflators (inc. GDP deflator), real wages continue to rise until 1936

Question 15

Devaluation recovery channels: Channel 2: International Export Competitiveness (MAYBE) (7) (RBCKS)

Answer 15

• By stimulating exports, devaluation stimulates aggregate demand
• Redmond (1980)
  o Calculates effective exchange rate (EER) for GB – weighted avg. of £ vs. all currencies
  o Shows that £ remained below August 1931 level through 1936 – LR benefits
  o £ appreciates from 1933 onwards, but only slowly
• Broadberry (1986) – uses elasticity approach to measure impact
  o Takes Redmond (1980)’s figures of 13% EER depreciation 1931-32
  o Assumes Marshall-Lerner condition holds
  o BoT improves by £80m, and assuming multiplier of 1.75, GDP effect is 3% overall
• Chadha et al. (2023) - estimate pass-through coefficient of 0.6 suggesting substantial (but slightly smaller) effect than Broadberry (1986)
• Kitson & Solomou (1990) – ½ of GB trade is with countries tied to sterling, so impact smaller
• Solomou (1996)
  o Estimates regression for exports (1932-37) – coefficient on ER regime is statistically insignificant (t=0.85) – exports det. by protectionism & trading blocs
  o Initial advantage eroded due to prices & devaluation of Gold Bloc
  o Devaluation only had an impact effect (32-33), but no LR effect (eroded by inflation)
  o UK – volume of exports in 1937 only 80% of the 1929 level
  o Trade hysteresis evident – M shares fell permanently despite higher ER later
• Overall, effect appears to be quite weak
  o Devaluers’ competitive edge gradually eroded by increased inflation
  o Trade increasingly conducted in trading blocs with similar ER strategies
  o Gold Bloc itself was devaluing
  o Global contraction meant there was insufficient world demand

Question 16

Devaluation recovery channels: Channel 3: Profitability (YES) (3)

Answer 16

• Eichengreen & Sachs (1985)
  o Higher competiveness (via devaluation) leads to higher profit, so more investment
  o Also stimulates investment via lower r and higher NPV of future profits
  o Tobin’s q=Security Prices/Output Prices - higher value means more incentive to invest
  o Graph shows clear negative relationship between ER and Tobin’s q – stimulated I
• Solomou & Weale (1997) – evidence from the UK supports this
  o Construct UK wealth figures for 1920-1956
  o Strong stock market boom in the UK between 1931 and 1936 – 100% increase in value
  o Net personal wealth increases a lot – 46% rise 1931-1936, driven by equity holdings
  o Wealth to income ratio up from 5 to 6 over this period too - higher consumer confidence
• Could also be attributed to ‘cheap money’?

Question 17

Devaluation recovery channels: Channel 4: MP Policy Freedom (YES) (4) EMKB

Answer 17

• Eichengreen (1991) – devaluation is ‘necessary precondition’ to policies for recovery
  o External constraint was binding – Gold Bloc M1 still 11% below 1929 level in 1935, ct 7% rise in M1 for sterling area by 1935
  o Removing the ‘external constraint’ leads to MP flexibility
• Morys (2014) – devaluation allowed GB to avoid banking crises
  o Stopping banking crises not feasible under Gold European crises (summer 1931)
  o In GB, BoE could be lender of last resort, ameliorating severity of Depression
  o Bernanke & James (1991) blame banking crisis for severity of US Depression
• Kitson & Solomou (1990) find this key – effect via freeing up MP, not ER itself
• Broadberry (1986) - Mundell-Fleming Model MP works better under flexible ER regime

Question 18

Monetary Policy (6) RBWK

Answer 18

• Takes a while to kick in because of initial fears of inflation having left Gold Standard
• Bank Rate reduced from 6% in February 1932 to 2% in June 1932 – “cheap money”
• Richardson (1967) – MP a ‘permissive’ but not ‘causal factor’
  o Bank Rate reduced to 2.5% in early 1931 but crisis in summer forced it back up (to attract gold reserves), so MP only works with good underlying conditions
  o Security prices up 50% between 1932 and 1933
  o Investment effect not key since firms rely increasing on profits for finance in 1930s – main effect of “cheap money” is allowing them to sell securities at profit to banks
• Broadberry (1986) – effectiveness of MP depends on
  o Interest elasticity of Md – seems to be low (0.2), so steep LM MP effective
  o Stability of Md – V fell sharply 1929-32 offsets expansionary MP
  o Interest elasticity of I – seems to be low (0.1-1.1), so steep IS MP ineffective
  o Exchange rate effect – clearly has to be key – estimates ER depreciation resulting from “cheap money” policy as increasing GDP by 3%
• Worswick (1984) – investment didn’t increase until after 1933, so may have been an accelerator effect rather than “cheap money”
• No convincing transmission mechanism (Kitson & Solomou, 1990)
  o There is no increase in bank advances until 1935, so must be via C/housing
  o However, MP doesn’t appear to be significant for housing
  o Lund & Holden (1968) – non-residential investment is interest inelastic
  o Direct effects are weak, but indirect effects stronger (profits, stock mkt, repayments)

Question 19

The Tariffs - timeline (5)

Answer 19

• 1915 – McKenna duties – 33.3% ad valorem duties imposed on luxury items (to save shipping space during WW1)
• 1921 – Safeguarding of Industries Act – duties on items crucial to national security/industry
• 1931 – only 2-3% of imports subject to tariffs
• November 1931 – Abnormal Importations Act (power to put 100% tariffs on any good)
• February 1932 – Import Duties Act
  o 10% ad valorem General Tariff - by mid-1930s, approx. 25% of imports subject to this rate
  o Empire was exempted from the General Tariff
  o Import Duties Advisory Committee (from April 1932) considered higher duties – most industries received extra protection (20%) and a few got even higher (30%)

Question 20

1920s Tariffs - motor industry (Foreman-Peck, 1979) (5)

Answer 20

• For private cars, additional 55% protection brought total effective protection to 88%, and 45% for commercial vehicles
• By 1924, vehicle yearly output 155,800 compared to 25,200 in 1912
• Morris (vehicles designed specifically for British market) surpassed British Ford for the first time with 22% of industry production
• Could argue this was simply due to a boom in the car manufacturing industry
  o But 1924 also saw election of minority Labour government which repealed McKenna duties
  o As a result, imports doubled and home-produced sales to domestic market virtually stagnated, until Duties reimposed following year
• Motor industry production exceeded 500,000 in 1937

Question 21

The General Tariff (9) MACECRBST

Answer 21

• Momentum behind protectionism (Keynes/Tories)
• Adopted to improve BoT and prevent excessive depreciation of £
• Classical theory promoting free trade leading to specialisation by comparative advantage relies on small open economy (no market power), full employment & no increasing RTS – very restrictive assumptions which may not be satisfied (Solomou, 1996)
• Economic theory of tariffs
  o Theory of Second Best – if there are already distortions (i.e. foreign tariffs), then one more distortionary policy may increase global welfare
  o Optimum Tariff – if a big country (i.e. can affect prices), imposing tariffs may mean that positive ToT effect (reduce M price) offsets negative welfare effect
  o Increasing returns – protection can create EoS in domestic industry
  o Macro effects – (M-F model) – under flexible ER, no effect on Y (but managed float) – but relies on assumptions of full employment, constant RTS & no uncertainty on I
• Clemens & Williamson (2004)
  o tariffs → ↑ growth relationship pre-1950 due to high level of tariffs elsewhere
  o Regressions show that tariff-growth relationship returns post-1950 if average tariff rates increase to pre-1950 levels (a kind of Prisoner’s Dilemma in tariffs pre-1950)
• Richardson (1967) sees main role of commercial policy as 1915/1921 tariffs
  o New industries key (“new development block”), but mostly protected by 1932
  o They were helped by earlier tariffs, creating EoS (e.g. synthetic dyes achieved EoS after 1921 tariffs) and then drove recovery – earlier tariffs are the crucial part
  o Revenue creation for government not key – only c.30% of customs revenue
• Broadberry (1986) – use IS-LM analysis
  o Under a floating ER, tariff leads to improvement on BoT, putting upward pressure on the ER, offsetting original effect on BoT – net effect on national income is zero
  o Tariff has no additional impact – just a substitute for further ER depreciation
• Solomou (1996) – trade hysteresis effects means that impact effect is important – gain market share & achieve EoS, generating a LR competitive advantage
• Timescale affects our verdict – LR effect unlikely to be positive
  o Likely to hold back structural change (see Crafts (2012)) – featherbeds old industries
  o Influences trade patterns – Empire vs. Europe

Question 22

General Tariff - Channel 1 - Import Substitution (6) (RKSBL)

Answer 22

• Import substitution is an economic policy aimed at reducing a country’s reliance on imports by promoting the development and production of domestic industry. This policy typically involves imposing tariffs or quotas on imported goods to protect domestic industries from foreign competition.
• Richardson (1967) – import replacement ratio shows tariff unimportant
  o Import replacement ratio IRR=∆Y-∆X/-∆M, calculates for each industry (30-35)
  o Theoretical norm is 1 for import substitution
  o If high, means that growth occurred independently of the tariff (not due to ∆M)
  o IRR is 50% higher (3 vs. 2) for newly protected industries, so tariff relatively unimportant in explaining 1930s resource flows
  o Problems: uses 1930/1935 Censuses of Production, but relevant measure is difference between 24-30 and 30-35
• Kitson & Solomou (1990) – tariffs play a key role, but are ‘incomplete explanation’ alone
  o Look at 1924-30 and 1930-35 growth rates of newly vs. non-newly protected
  o Inter-period output growth difference is 4pp for newly protected vs. -0.4pp for non-newly protected (control group), which is statistically significant tariff is key
  o Tariff is 2nd best response to foreign tariffs (US – Smoot-Hawley (1930) – 45-50%)
• Solomou (1996) – calculates import demand function, M=a+b1Y+b2P+b3t, finding t variable significant at 99% significance level – 1% increase in t reduces M by 3.4%, which is greater than b2 (price effect), so tariff also works through other channels (confidence etc.)
• Broadberry & Crafts (2011) – tariff not good/useful for recovery
  o Innovate on K&S (1990) by using formal difference-in-difference methodology and by using 3 different groups (early protected, additional duties & others)
  o Run regression 1930/35-1924/30 =a+bAdditional+cEarly (Y/L used)
  o None are very significant (b is at 20% – so any effect via additional duties)
  o Suggests that tariff had little effect on recovery in the 1930s
• Lloyd-Solomou (2020) criticisms of Broadberry & Crafts (2011)
  o Restrict “early” protected to 15 industries, meaning it is unrepresentative – shows Y/L growing at 4% p.a. 1924-30, growth rates not even the US reaches
  o “Others” have either 0% or 10% tariffs – but 0% shouldn’t be in this group (they are non-policy intervention) – should either use them as control or get rid of them
  o Using difference-in-difference model & Kitson-Solomou data set, output and productivity effects are very significant (5% level) – no classification issues (2 groups)
  o Modifying Broadberry-Crafts sample to have new control as non-newly protected shows that “additional” group is significant at 5% level (both growth & productivity) – this is a substantial group, so tariff has significant effect overall

Question 23

General Tariff - Channel 2 - Effective Protection Rates (INCONCLUSIVE) (4)

Answer 23

Effective rate of protection recognises that an industry’s net benefit depends on how the cost of its inputs change – measures degree of protection on value added, rather than on product price.

• Capie (1978) looks at effective protection rates
  o Provisos: assumes infinite supply elasticity from rest of world (so overestimating protection)
  1. Case studies
  o Little effective protection for iron/steel (least in manu.), yet seen as driving recovery
  o Negative effective protection for building (0% tariff and 42% of inputs are imported), yet also seen as a key growth sector
  o Clearly, growth sectors in 1930s are not driven by the tariff
  2. All industries
  o Calculates rank correlation coefficient between effective protection and growth
  o Yields an inverse relationship –> protectionism is holding back the recovery
• Foreman-Peck (1981) is critical of Capie (1978)
  o Capie assumes that the prices of inputs increase by the full amount of the tariff – requires perfect substitutability of imports for home products – unlikely
  o Calculates these figures (IRR framework), showing that iron/steel benefitted a lot & little impact on building (costs don’t rise)
• Kitson, Solomou & Weale (1991)
  o Recalculate effective protection rates, doing it properly (however, still assuming that prices increase by the full amount of the tariff)
  o Case studies are not robust
   Building still negative (nontradeable), but tariff may have had +ve Y effect
   Iron & steel go to top 10% of results
  o Correlation not robust either – no relationship between growth and EPR (careful to interpret – doesn’t tell us about impact of tariffs since industries are ones protected in 1921 and 1932, so says nothing about the 1930s policy intervention)
  o EPRs have highly restrictive assumptions (full employment, perfect competition, constant RTS & law of one price)
  o Qualitative policy shift is key – 80% of UK industry protected for the first time (1932)
• In conclusion, effective protection rates don’t give a decent conclusion on the debate

Question 24

General Tariff - Channel 3 - Macroeconomic Analysis (MAYBE) (2)

Answer 24

• Foreman-Peck (1981)
  o Imports of manufacturers fall by 32.7% (30-35)
  o Concludes that PED for imports is 2.1 (adjusting for Y growth); with multiplier of 1.75, this means GNP up 4.1%
  o Broadberry (1986) finds problems with this – firstly, assumes constant ER (30-35) based on £/$, but EER is not constant (depreciated, so would decrease imports); also ignores effect on ER of tariff (appreciated, so would increase imports) – both reduce the effect of the tariff on national income
• Eichengreen (1979) – general equilibrium effects may be much smaller
  o Tariff increased domestic income & impact on asset markets raised ER – partially offsets the expenditure switching effects
  o Estimates S equation model & simulates removal of tariffs
  o Initial effects were small; GDP 2.3% higher as result of tariff by 1938
  o Broadberry (1986) finds problems with this too – very low R2 & autocorrelation too

Question 25

General Tariff - Channel 4 - Trade Blocs/Imperial Preference (MAYBE) (3)

Answer 25

• 1932 Ottawa Agreements – free trade maintained among sterling countries
• Sterling area also emerged – countries sharing common currency ( trade via ER stability)
• Eichengreen & Irwin (1995)
  o £ area traded unusually extensively internally (ER stability) & externally (deval)
  o Use gravity model of trade to investigate the impact of trading blocs – includes distance & contiguity (share border)
  o Dummy for Commonwealth shows higher trade in bloc 1932 to 1935 (significant), so trade bloc was trade creating (only small diversion effects, not always significant)
  o Part of this effect is visible in 20s (existing links), but increases in 30s
  o Differential success of trading blocs – German one (E Europe) not too successful
  o No +ve effect on trade within sterling area – ER stability not very important for trade

Question 26

Fiscal Policy - Traditional View (3)

Answer 26

• Richardson – ‘the government did not have a fiscal policy as such in the 1930s’. Broadly neutral in the early part of 1930s.
  o Historians tended to look at the actual budget balance.
  o Once fiscal window-dressing (statistical artefacts to help governments look as if they have balanced the books) removed, the budget balance was broadly zero. Therefore not really significant for growth.
  o Based on nominal budget surplus, policy was slightly expansionary 1930-31 to 1932-3 and was slightly contractionary until rearmament, suggesting a counter-cyclical stance.
  o Balanced budgets of 1930s created impression that fiscal policy was neutral.
• Broadberry (1986) – there are big problems with just looking at nominal budget surplus
  o Different items affect economy in different ways, so must look at fiscal leverage
  o Changes in prices mean changes in real debt payments, changing ‘real’ surplus

Question 27

Fiscal Policy - Modern View (11) MBCMTBCMRK

Answer 27

Must look at Constant Employment Budget Balance (CEBB)
Bs=T-G
T=tY
∴Bs=tY-G=tY* (Y/Y*) - G
where Y* is the full employment level of Y; assuming full employment,
CEBB=tY*-G
This can only vary via discretionary changes (can also look at it as percentage of Y*)

• Middleton (1981)
  o Looks at the Constant Employment Budget Balance (CEBB) – Bs doesn’t distinguish between discretionary changes and autonomous changes in economic activity
  o A cyclical downturn means that ex post Bs falls unless increase t or reduce G
  o Therefore, balanced budget suggests a heavily deflationary stance (esp. 1931)
  o Manipulated receipts over time from funds (“window dressing”) to ensure a balanced budget – evidence of this in early 1930s (gets rid of this in CEBB figures)
  o Uses Y* as trend from 1929-30 to 1937-38
  o 1931 budget saw increase in CEBB surplus by 1.4% – very deflationary
  o CEBB was deflationary by -3% GDP in 1932-33 and -4% GDP in 1933-34
  o Loosened by 1% GDP p.a. 1934-38, so sustained recovery thereafter
  o Thus there was a structural surplus at the trough of the recession, structural deficit once it recovered (but only through rearmament, at the end). Destabilising.
• Broadberry (1984) – changes in fiscal leverage don’t show the same destabilising pattern
  o Middleton doesn’t take into account price changes - falling prices in early years of depression would increase the real value of interest payments, boosting demand
  o Uses ‘fiscal leverage’ – once this measure is used, the strongly destabilising influence identified by Middleton is not found.
  o Use a model to describe differences in GDP if there had been no govt sector at all
  o Also includes Pigouvian real balance effect – lower prices mean higher real interest payments to households, offsetting deflationary impact of CEBB
  o On this basis, fiscal policy is neutral 1930-31 to 1933-34 (tax multiplier)
  o N.B. Broadberry doesn’t argue that the budget wasn’t expansionary in later years especially in 1937-8 (agrees with Middleton)
• Cloyne et al. (2023) – use narrative evidence to identify exogenous changes to taxation and so determine tax multiplier
  o Important since most fiscal consolidation through tax
  o Find tax multiplier = 2-3 within 2 years of change
  o Suggests higher taxation in early 30s was very contractionary
• Middleton (1984) is sceptical
  o Doesn’t like fiscal leverage concept – doesn’t distinguish between discretionary and automatic components of fiscal policy – defeats the point of CEBB
  o Real balance effect unlikely – most debt held by institutions, given recession, propensity to save is high, propensity to invest is low 🡪 small effect on AD
• Turner (1991) accepts fiscal leverage concept but rejects real balance effect (consumption elasticities on financial assets are small), concludes that fiscal policy was broadly neutral across recession
• Broadberry (1986) – IS/LM/BP analysis shows that fiscal policy works much better with fixed ER regime, so perhaps role is not so key in 1930s
• Crafts & Mills (2012) – fiscal stimulus would not have been effective
  o Use “defence news” to capture exogenous shocks to key component of G (1932-38) (unrelated to cycle)
  o Calculate a multiplier, getting range of 0.3-0.8 (ct historiography – 1.2 to 2.5)
  o Fiscal stimulus would have been ineffective – low multiplier & possible increase in risk premia leading to the end of “cheap money”.
• Middleton (1984) – fiscal policy did help by initiating “cheap money”
  o Increased business confidence from adherence to balanced budget (also increases investment)
  o Increased savings from lower spending 🡪 lead to lower real interest rates
  o War Loan conversion to reduce debt charge and maintain balanced budget.
• Some economists argue 1931 budget was neutral or even positive for the economy (via confidence):
  o Less risk of a government default, therefore lower risk premium on bonds, so interest rates lower, encouraging investment.
  o Consumers are also more confident about the outlook (sound economy), so spend money.
  o Middleton rejects this because the effects of higher confidence would have to be huge to offset such a fiscal contraction.
• Richardson (1967) – 1931 balanced budget had a significant, positive confidence effect
  o All contemporaries agreed that autumn 1931 crisis budget had +ve effect on I
  o Problem was pessimism of business, so increasing confidence allowed recovery
  o The Economist rejected public works programme in 1932 due to confidence effect
  o Acknowledges that confidence probably didn’t compensate for real influences
• Kitson & Solomou (1990) – confidence effect would have to be implausibly large to offset the heavily deflationary stance from the CEBB

Question 28

Fiscal Policy - Rearmament (7) CRTC(RC)C

Answer 28

• Crafts & Mills (2012) – had a substantial impact by late 1930s - responsible for 30% of GDP increase between 1932 and 1938 – mainly after 1935
• Richardson (1967) – strongly counter-cyclical in 1937-38 recession
  o 1938-39 defence spending was 4x level of 1935-36 (30% of total G by then)
  o Offset negative effects of 1937-38 recession – u/e only increased by 2.5% and recovery occurred in January 1939, due to rearmament spending

Thomas (1984) – uses a Social Accounting matrix with input-output tables to estimate multiplier
o Estimates multiplier of 1.6 1935-1938 🡪 increase in GDP by 6.5% from 1934-38
o Iron, coal, steel, engineering benefitted with large employment growth.
o Rearmament > housing boom since the latter was biased towards Southern England, whereas former more even regional distribution of effect

• Crafts (2013) – plausible that rearmament boosted recovery post-1935
  o End of 1935 - defence spending up 28% compared to previous year
  o Feb 1935 – Defence White Paper warned of £1500 million spending over next 5 years (partly deficit financed); Defence Loans Bill approved - authorised £400mn borrowing over 5 years
  o Believed that this promoted substantial GDP and employment increases

Anticipation effects
o Robertson (1983) - News of rearmament stimulating growth in construction industry as armaments manufacturers increase production capacity.
o Chambers (2010) – shows with IPO data, increase in aircraft manufacture ventures from 1934-36, despite relatively small increase in defence expenditure

• Crafts and Mills (2013)
  o “rearmament delivered a valuable, if limited, stimulus after 1935”
  Low multiplier (0.3-0.8) due to 1. expectation of higher future taxation and 2. Rearmament investment is not efficient, since it doesn’t produce longer-term growth
  o Estimate that, in the absence of rearmament, real GDP in 1938 would have been at most 5% lower

Question 29

Policy Regime Change - definition (2)

Answer 29

• Sargent (1983) – underlying policy regime is the systematic and predictable part of all decisions – the thread that runs through the individual choices that govts/CBs have to make
• New regime has to be a dramatic change that is clearly articulated and well understood
  1930s regime change - pushed market price expectations upwards (low interest rates, devaluation tariff - government appears credibly committed)

Question 30

Policy Regime Change - timeline (5)

Answer 30

1. Macmillan Report, published 13 July 1931 - the first quasi-official declaration to raise prices
   * ‘A large rise towards the price level of 1928 is greatly to be desired… We recommend that this objective be accepted as the guiding aim of the monetary policy of this country’
2. Leaving the gold standard (21st September 1931) (forced, not choice) – changed policy trilemma faced by policymakers - allowed potential use of discretionary MP to respond to depression.
3. Implementation of cheap money policy - saw Bank rate cut in steps from 6% to 2% (Feb 1932 - June 1932), 2% maintained until Aug 1939 – rise in price level promoted through cheap money, weak pound – March 1932 $3.40 (30% devaluation from $4.86)
4. Import Duties Act - 10% tariff levied on many imported goods (March 1932) - encouraging firms to exploit enhanced market power
5. Chancellor embraced recommendations of Macmillan Report, announces objective of raising prices back to 1928 level at British Empire Economic Conference in Ottawa in July 1932
   * ‘we have got to raise wholesale prices’ – emphasis on wholesale vs retail prices due to desire to increase profitability of British industry to restore confidence
6. War Loan Conversion - perpetual bonds converted from 5% to 3.5% - signalling effects + frees up G (less debt interest payments)

Question 31

Policy Regime Change - (Temin, 1989) (8)

Answer 31

• Draws on Sargent (1983) to emphasise importance of expectation changes as part of an underlying policy regime change to help economies out of Great Depression
• Unlike Germany and USA, there was no policy regime change in the UK because changes were
  too piecemeal. 3 main factors:
  o Changes in the UK were less radical, especially with fiscal policy. Middleton – CEBB peaked in 1933/4 (+4.2%), so fiscal policy was getting tighter and tighter. Therefore, it was subtracting from growth. Hard to believe that there was a newfound desire to do whatever it took to get economy growing.
  o Changes in the UK’s policies were not chosen deliberately, but rather were forced upon them. If not done willingly, then are not credible because they may be reversed in the future, once the government is strong enough to reverse them. Britain’s devaluation in September 1931 was a one-off (due to speculation), monetary policy changed due to government debt worries and fiscal policy was contractionary.
  o Timing of Recovery – According to rational expectations, recovery should be immediately after the policy regime change has occurred. Rational actors should recognise it immediately and respond immediately. The lag was too long (recovery only starts in winter of 1932) to be consistent with this theory.
• Hence GB did not see a policy regime change, preventing effective recovery
• Ct USA – devaluation in March 1933, perceived as policy regime change –> recovery
• US is much more immediate – devalue under no pressure, New Deal & MP in 1933:
  o Stock prices doubled in 2nd quarter of 1933 (devaluation in April).
  o Cotton price doubled too (compared to no reaction in UK).
  o Investment doubled within a few months.
  o Clear, dramatic, consistent regime change.
• UK recovery due to external stimulus & expansionary MP:
  o Forced off gold, MP due to fiscal needs, tariff due to retaliation vs. US Smoot Hawley Tariff, fiscal similar – not decisive/consistent.
  o Commodity prices don’t respond in the UK.
• Devaluation is the best indicator of regime change (tangible) – so Eichengreen & Sachs (1985) are right, but mechanism is expectations

Question 32

Policy Regime Change - (Mitchell et al., 2012) - high frequency (monthly) UK data (4)

Answer 32

• Devaluation in September 1931 correlated with significant recovery (small expectations effect since immediate trade response unlikely), but recovery came to abrupt end in early 1932 – suggests limited policy regime change
• Economy continued to slide between February and September 1932, despite General Tariff (February) & beginning of “cheap money” (February to June)
• Consistent with Temin’s hypothesis ASSUMING RATIONAL EXPECTATIONS – no observable recovery in GB 1931-32
• Recovery began in October 1932 – remains weak in 1933 but thereafter, a persistent & strong recovery

Question 33

Policy Regime Change - (Mitchell et al., 2012) - rational expectations vs adaptive learning (5)

Answer 33

• Temin’s evaluation of UK experience is based on rational expectations framework
• Rational expectations’ assumptions are unrealistic – perfect knowledge, unique model for economy & knowledge of its parameters. Therefore, they argue with more relaxed assumptions there was a policy regime change
  Why?
  o British economic performance was very strong – 4% per annum, given it had not had a severe depression compared to the rest of the world.
  o The lack of an expansionary fiscal policy was defensible and consistent with a policy regime change – balanced budget was needed to maintain confidence in the economy, for investment.
  o Rational expectation is too restrictive. People do not have perfect information, time lags are expected (see below) etc.
• Bray & Savin (1986) suggest that the learning path of expectations is more important
• Adaptive learning approach – agents have to take time to gather info and learn about parameters of model – takes time to get to rational expectations equilibrium
• Evidence from high frequency data is consistent with idea that it took time to learn that equilibrium expectations had changed
• Ellison et al. (2024) - use a nowcasting methodology to estimate inflation expectations for 27 countries.
  o Find increase in expectations began in 1931 (after devaluation), settled before end of 1932, and then increased again afterwards 🡪 Temin wrong.
  o Consistent with Mitchell et al. (2012) adaptive learning model

Question 34

Policy Regime Change - (Mitchell et al., 2012) - adaptive learning view fits other evidence (2)

Answer 34

• Eichengreen & Sachs (1985) – on average, there was a 2 year lag between devaluation in the early 1930s and monetary expansion
• Ritschl (2002) – German high frequency data suggests recovery is built over a longer phase between 1932 and 1933 – hence it was learning process

Question 35

Was policy regime change consistent? (2)

Answer 35

• GB can be seen as a consistent policy regime change
  o Lag to adopt “cheap money” (less than 1 year) was less than average (2 years)
  o Fiscal policy is consistent – confidence effect (1931) & small multiplier (Crafts & Mills, 2012), so expansion would have had little effect anyway (although questionable that the govt knew this)
• In this view, recovery occurs due to devaluation (with a lag) and subsequent shift to “cheap money” policies, possibly including the General Tariff too

Question 36

Policy Regime Change - (Temin, 1989) - international comparison (US, UK, Japan, Argentina) (4)

Answer 36

• US is much more immediate – devalue under no pressure, New Deal & MP in Apr ‘33
  o Stock prices doubled in 2nd quarter of 1933 (devaluation in April)
  o Cotton price doubled too (ct no reaction in UK)
  o Investment doubled within a few months
  o Clear, dramatic, consistent regime change
• UK recovery due to external stimulus & expansionary MP
  o Forced off gold, MP due to fiscal needs, tariff due to retaliation vs. US’ Smoot Hawley Taruff (cf Canada), fiscal similar – not decisive/consistent
  o Only consistent if know that fiscal multiplier < 1 – difficult to know
  o Commodity prices don’t respond in the UK
• Japan – “Takahasi Public Finance”
  o Japan outperforms, given ER movements – devaluation not full explanation
  o Huge fiscal stimulus – G becomes largest part of AD; MP expansionary; W restraint
  o Changed expectations immediately
• Argentina – clear evidence of regime change
  o Prebisch reduces interest rate, introduces exchange controls, devalues etc.
  o Della Paolera & Taylor (1999) use dynamic model to see influence without policies of conversion office/Central Bank – real interest rates 20-30% lower than otherwise – hence, recovery based on regime change of 1931 – via expectations (Mundell Effect)

Question 37

Policy regime change - Lennard et al. (2023) - Quantitative news (+ criticism) (1 + 3 + 1)

Answer 37

• Count number of articles containing stems ‘inflat’ and ‘deflat’ based on Binder’s idea that “media coverage reflects and shapes macroeconomic expectations of the public”
• Data shows 3 major discontinuities coinciding almost precisely with significant policy decisions:
  1. 21st September 1931 (gold standard abandonment) – spike in inflation reporting
  2. January 1933 – articles discuss prospect of global inflation
  3. April 1933 onwards – sustained increase in inflation reporting after dollar devaluation
• However – only stems accounted for, not focus of article, difficulties in discerning between national and foreign news

Question 38

Policy regime change - Lennard et al. (2023) - Commodity prices (+ criticism) (3)

Answer 38

(See notes for equation)
* Observable LHS – used to estimate expected inflation
* However – data for spot & future prices for copper, cotton, tin – on Board of Trade wholesale prices index, these commodities allocated 5%, 2%, 1% weight, so only make up a fraction of aggregate price index and not necessarily correlated with price movements for other goods
* Unrealistic assumptions – agents risk-neutral, market neither segmented nor underdeveloped, lack of storage costs

Question 39

Policy regime change - Lennard et al. (2023) - Findings (weighted principal components analysis with bond term premia) (4)

Answer 39

• Inflationary expectations indeed positive from end of 1930, spikes in inflationary expectations coinciding with important policy announcements
• Positive correlation with previous analysis – 0.38 with Hamilton commodity price estimates, 0.79 with Binder quantitative news estimates (US) – suggests interconnectedness of global economy – major international shocks (France, Germany, US) certainly shaped UK inflationary expectations
• Possibility of reverse causality – expectations –> economic outcomes, or outcomes –> expectations? Was price recovery inevitable given Britain’s strong growth? Perhaps transmission mechanism of MP + government credible commitment indicates former
• No high frequency data for components of GDP – use VAR model
  o Suggests deflationary expectations hindered growth until November 1931
  o Not until April 1933 that inflationary expectations became sustained growth stimulus – from this point until January 1934 (when GDP exceeded its pre-recession peak) model suggests that inflationary expectations accounted for 69% of the 5.6% increase in real GDP
  o Strongly indicates the existence of an important channel through which inflationary expectations alleviated 1930s slump, ascribes an increasingly important role for inflationary expectations, perhaps as more policies were added to reinforce the policy framework and in response to the US devaluation

Question 40

1930s Economic Recovery - Other (2)

Answer 40

• Crafts (2012) – weakness of competition in Britain in late 1930s undermined productivity growth from 1930s through to 1970s
  o Theoretically, competition improves productivity, increases the adoption of new technology (Schumpeter’s creative destruction) & reduces TU power
  o Empirically, via all 3 channels, competition seems to enhance productivity too
  o 1930s saw retreat from competition – abandoning Gold, General Tariff & more industrial policy
  o By 1935, at least 29% of manufacturing output was cartelised
  o Comparing 1924/1935 Censuses of Production, price-cost margin increases by 3.8%
  o TFP fell below 1924-1929 rate in 1929-1937 – due to poor competition
• Kitson & Solomou (1990) – interaction of trade & natural causes is important, but overall, trade is the key driver of strong growth in the 1930s