1What is the production function in the task-based automation model?
1. $Y(t)=\widetilde Z(t)\prod_{i=1}^{N}X_i(t)^{1/N}$.
2. $X_i(t)$ is the service supplied by task $i$, with $N$ equally weighted tasks.
2How are tasks performed in the task-based automation model?
1. Non-automatable: $X_i(t)=L_i(t)$.
2. Automatable: $X_i(t)=L_i(t)+K_i(t)$.
3What is the distinction between substitution within tasks and complementarity across tasks?
1. Within automatable tasks, capital can substitute for labour.
2. Across tasks, different task services must be combined to produce output.
4Why can the task-based model maximise $\ln Y(t)$ instead of $Y(t)$?
1. $\ln Y(t)$ is strictly increasing in $Y(t)$.
2. A strictly increasing transformation preserves the maximiser while simplifying the objective.
5How is the extent of automation defined in the task-based model?
1. $\alpha=m/N$, where $m$ is the number of automatable tasks and $N$ is the total number of tasks.
6When do capital and labour specialise across tasks?
1. If $w(t)>r^K(t)$, then capital is cheaper for automatable tasks.
2. $\forall i\leq m$, $X_i(t)=K_i(t)$; $\forall i> m$, $X_i(t)=L_i(t)$.
3. This requires verification using equilibrium factor prices.
7What are the optimality conditions for allocating inputs across tasks in the automation model?
1. Capital: $\frac{1}{NK_i(t)}=\frac{1}{NK_j(t)}$, $\forall i,j\leq m$.
2. Labour: $\frac{1}{NL_i(t)}=\frac{1}{NL_j(t)}$, $\forall i,j>m$.
8How are capital and labour allocated across tasks in the automation model?
1. $\forall i\leq m$, $K_i(t)=\frac{K(t)}{m}$.
2. $\forall i>m$, $L_i(t)=\frac{L(t)}{N-m}$.
9How is aggregate production derived from the optimal task allocation?
1. $Y(t)=\widetilde Z(t)\prod_{i=1}^{N}X_i(t)^{1/N}$.
2. $Y(t)=\widetilde Z(t)\left(\frac{K(t)}{m}\right)^{m/N}\left(\frac{L(t)}{N-m}\right)^{(N-m)/N}$.
3. Using $\alpha=m/N$: $Y(t)=Z(\alpha)K(t)^\alpha L(t)^{1-\alpha}$.
10How does automation affect aggregate production in the task-based model?
1. $\alpha=m/N$: the fraction of automatable tasks.
2. $Z(\alpha)=\frac{\widetilde Z/N}{\alpha^\alpha(1-\alpha)^{1-\alpha}}$.
3. Automation changes both the capital exponent $\alpha$ and TFP $Z(\alpha)$.
11How does the automation share determine factor income shares?
1. Capital share: $\frac{r^K(t)K(t)}{Y(t)}=\alpha$.
2. Labour share: $\frac{w(t)L(t)}{Y(t)}=1-\alpha$.
3. $\alpha=m/N$ is the proportion of automated tasks.
12What condition supports complete specialisation in the automation model?
1. $w(t)>r^K(t)$: capital is cheaper than labour.
2. $\frac{w(t)}{r^K(t)}=\frac{(1-\alpha)K(t)}{\alpha L(t)}$.
3. Therefore, $\frac{K(t)}{L(t)}>\frac{\alpha}{1-\alpha}$.
13How does automation affect output in the task-based model?
1. $\frac{\partial\ln Y(t)}{\partial\alpha}=\ln\left(\frac{(1-\alpha)K(t)}{\alpha L(t)}\right)$.
2. $=\ln\left(\frac{w(t)}{r^K(t)}\right)>0$.
3. Automation increases output by replacing expensive labour with cheaper capital.
14How does automation affect total labour income?
1. $w(t)L(t)=(1-\alpha)Y(t)$.
2. $\frac{\partial[w(t)L(t)]}{\partial\alpha}=(1-\alpha)\frac{\partial Y(t)}{\partial\alpha}-Y(t)$.
3. The result follows from the product rule.
15What are the productivity and displacement effects of automation?
1. Productivity effect (first term): automation increases output, benefiting labour.
2. Displacement effect (second term): automation decreases labour's share of output.
3. The net effect on labour income is ambiguous.
16How does automation affect wages in the task-based model?
1. $\frac{\partial\ln w(t)}{\partial\alpha}=-\frac{1}{1-\alpha}+\frac{\partial\ln Y(t)}{\partial\alpha}$.
2. $=\ln\left(\frac{(1-\alpha)K(t)}{\alpha L(t)}\right)-\frac{1}{1-\alpha}$.
3. Wages increase if the productivity effect exceeds the displacement effect.
17How does capital abundance affect the wage response to automation?
1. Higher $K/L$ strengthens the productivity effect.
2. Wages increase if the productivity effect exceeds the displacement effect.
18What do Acemoglu and Restrepo (2022) find about task displacement and wages?
1. Greater task displacement is associated with weaker real wage growth across US demographic groups.
2. Evidence: 1980-2016.
19How do Acemoglu and Restrepo (2022) identify the wage effects of task displacement?
1. 2SLS across 500 US demographic groups (1980-2016).
2. Instrument displacement using initial industry and routine-task exposure interacted with subsequent automation adoption.
20What qualifies the causal interpretation of Acemoglu and Restrepo (2022)?
1. IV validity requires automation adoption to affect relative wages only through task displacement.
2. The negative estimate persists with additional controls ($\hat{\beta}=-1.56$).
21How are household income shares determined in the task-based automation model?
1. $I_h(t)=w(t)L_h(t)+r^K(t)K_h(t)$.
2. $\ell_h=L_h(t)/L(t)$, $\kappa_h=K_h(t)/K(t)$.
3. $I_h(t)/Y(t)=(1-\alpha)\ell_h+\alpha\kappa_h$.
22How does automation affect household income shares?
1. $\frac{\partial(I_h(t)/Y(t))}{\partial\alpha}=\kappa_h-\ell_h$.
2. The income share increases if $\kappa_h>\ell_h$.
23How does automation affect household income levels?
1. $\frac{\partial\ln I_h(t)}{\partial\alpha}=\frac{\partial\ln Y(t)}{\partial\alpha}+\frac{\kappa_h-\ell_h}{(1-\alpha)\ell_h+\alpha\kappa_h}$.
2. First term: aggregate output gain.
3. Second term: redistribution towards capital owners.
24What is the intuition behind the two effects of automation on household income?
1. Aggregate gain: automation increases total output, potentially benefiting all households.
2. Distributional change: automation shifts income towards capital, benefiting households with relatively more capital ownership.
25What did Meade (1964) warn about regarding automation?
1. Automation could concentrate income among wealthy capital owners while depressing wages.
2. He advocated broader private ownership and some public ownership of capital.
26Why might automation decrease the capital income share in richer task-based models?
1. Remaining human tasks can become increasingly valuable as machine services become abundant.
2. Their expenditure share can increase even as the number of human tasks decreases.
27What policies does Aghion (2025) recommend for inclusive innovation?
1. Competition: prevent concentration of AI inputs from restricting entry.
2. Opportunity: broaden access to education and innovation.
3. Flexicurity: combine income insurance, retraining and job-search support.
28What do Piketty and Zucman (2014) show about wealth-to-income ratios?
1. Private wealth increased relative to national income in many developed countries after 1970.
2. European ratios were also high before WWI, before decreasing and subsequently recovering.
3. Higher wealth-to-income ratios do not necessarily imply higher inequality.
29What is Piketty's first law of capitalism?
1. $\alpha_K(t)=r^K(t)\frac{K(t)}{Y(t)}$.
2. Capital's income share equals its return multiplied by the capital-to-income ratio.
3. An accounting identity.
30What saving assumption underlies Piketty's weak version?
1. A fixed gross saving rate $s$.
2. $\dot K(t)=sY(t)-\delta K(t)$.
31How is the capital-to-income ratio determined on the BGP under Piketty's weak version?
1. $g=s\frac{Y(t)}{K(t)}-\delta$.
2. $\frac{K(t)}{Y(t)}=\frac{s}{g+\delta}$.
3. With $s,\delta$ fixed, slower growth increases the ratio.
32What saving assumption underlies Piketty's strong version?
1. A fixed net saving rate $\tilde{s}$.
2. $\dot K(t)=\tilde{s}[Y(t)-\delta K(t)]$.
3. Saving is proportional to income after depreciation.
33How is the capital-to-income ratio determined on the BGP under Piketty's strong version?
1. $g=\tilde{s}\left[\frac{Y(t)}{K(t)}-\delta\right]$.
2. $\frac{K(t)}{Y(t)}=\frac{\tilde{s}}{g+\tilde{s}\delta}$.
3. With $\tilde{s},\delta$ fixed, slower growth increases the ratio.
34What is Piketty's second law of capitalism?
1. $\frac{K(t)}{Y(t)-\delta K(t)}=\frac{\tilde{s}}{g}$.
2. Capital relative to net income equals the net saving rate divided by aggregate growth.
35How does slower growth affect the gross saving rate under Piketty's strong version?
1. $s(g)=\frac{\tilde{s}(g+\delta)}{g+\tilde{s}\delta}$.
2. Slower growth increases the gross saving rate, unlike the weak version where $s$ is fixed.
3. More gross income is devoted to replacing depreciating capital.
36What happens as aggregate growth approaches zero under Piketty's weak and strong versions?
1. Weak: $\frac{K(t)}{Y(t)}\to\frac{s}{\delta}$.
2. Strong: $\frac{K(t)}{Y(t)}\to\frac{1}{\delta}$.
3. Strong: $\frac{K(t)}{Y(t)-\delta K(t)}\to\infty$ and $\frac{C(t)}{Y(t)}\to0$.
37What empirical evidence challenges the fixed-saving assumptions in Piketty's model?
1. Piketty and Zucman (2014): saving rates differ substantially across countries.
2. Krusell and Smith: US gross and net saving rates vary over time, with net saving declining markedly.
38What is the competitive rental rate of capital under CES production?
$r^K(t)=\frac{\partial Y(t)}{\partial K(t)}=\alpha\left(\frac{Y(t)}{K(t)}\right)^{1/\sigma}$.
39How is the capital income share determined under CES production?
1. $\alpha_K(t)=r^K(t)\frac{K(t)}{Y(t)}$.
2. $\alpha_K(t)=\alpha\left(\frac{K(t)}{Y(t)}\right)^{(\sigma-1)/\sigma}$.
40How does an increase in $K(t)/Y(t)$ affect the capital income share under CES production?
1. $\sigma>1$: increases.
2. $\sigma=1$: unchanged.
3. $0<\sigma<1$: decreases.
41Why does the elasticity of substitution determine how the capital share responds to an increase in $K(t)/Y(t)$?
1. $\sigma>1$: Capital is easily substituted for labour, so its return decreases less than proportionately, so its income share increases.
2. $\sigma<1$: Capital and labour are complements, so its return decreases more than proportionately, so its income share decreases.
42How is output per unit of capital expressed under CES production?
1. Divide the CES production function by $K(t)$.
2. $\frac{Y(t)}{K(t)}=\left[\alpha+(1-\alpha)\left(\frac{A(t)L(t)}{K(t)}\right)^{\frac{\sigma-1}{\sigma}}\right]^{\frac{\sigma}{\sigma-1}}$.
43What condition must CES production satisfy on the Solow BGP?
1. Capital accumulation requires $\frac{Y(t)}{K(t)}=\frac{g+\delta}{s}$.
2. The CES production function must support this ratio at a finite, positive $\frac{K(t)}{A(t)L(t)}$.
44What are the CES conditions for an interior BGP?
1. $\sigma>1$: $\frac{g+\delta}{s}>\alpha^{\frac{\sigma}{\sigma-1}}$.
2. $0<\sigma<1$: $0<\frac{g+\delta}{s}<\alpha^{\frac{\sigma}{\sigma-1}}$.
3. $\sigma=1$: $g+\delta>0$ suffices.
45How is the capital income share determined on the CES BGP?
1. $\alpha_K(t)=\alpha\left(\frac{K(t)}{Y(t)}\right)^{\frac{\sigma-1}{\sigma}}$.
2. Substitute $\frac{K(t)}{Y(t)}=\frac{s}{g+\delta}$.
3. $\alpha_K^*=\alpha\left(\frac{s}{g+\delta}\right)^{\frac{\sigma-1}{\sigma}}$.
46How can slower growth increase the capital income share in Piketty's CES model?
1. $\frac{K(t)}{Y(t)}=\frac{s}{g+\delta}$: slower growth increases $K/Y$.
2. If $\sigma>1$, the return on capital decreases less than proportionately.
3. Hence, $\alpha_K(t)=r^K(t)\frac{K(t)}{Y(t)}$ increases.
47What evidence challenges Piketty's assumption that $\sigma>1$?
1. Gechert et al. (2022) review 3,186 estimates from 121 studies.
2. Their corrected estimate is approximately $\sigma=0.3$, suggesting capital and labour are complements.
48How does the net capital income share differ from the gross share?
1. Gross: $\alpha_K(t)=\frac{r^K(t)K(t)}{Y(t)}$.
2. Net: $\alpha_K^{net}(t)=\frac{[r^K(t)-\delta]K(t)}{Y(t)-\delta K(t)}$.
3. Accounting for depreciation can change the response of the capital share.
49Why does the distinction between gross and net capital income shares matter for inequality?
1. A larger gross capital share need not mean capital owners receive a larger share of income available for consumption.
2. Depreciation absorbs part of capital income without increasing owners' purchasing power.
50How can income inequality encourage economic growth?
1. Effort and talent: Rewards encourage skill development and productive effort.
2. Entrepreneurship: Potential returns encourage risk-taking and innovation.
51How can income inequality harm economic growth?
1. Unequal opportunity: Restricted access to education and finance leaves talent underused.
2. Rent-seeking: Economic power can secure privileges and restrict competition.
3. Political polarisation: Income divisions can obstruct cooperation and reform.